ATI Paragon PPS1 Power Supply repairs

[nevetslab]

Now that I've technically moved out of the main CenterStaging Inventory warehouse and back into my small shop in the rehearsal studio complex, and have updated the desktop computer enough to get billing done and move onto the next project, still fighting for bench space, that next project is a pair of 70+ lb ATI Paragon PPS1 Power Supplies that run those beautiful live sound consoles. Winn Krozack runs the Paul Reed Smith Artist Relations office across the street in our Industry A & R offices, and until recently, he also had the Eagles custom console that Cal Perkins designed for them from the mid-70's there, along with one of these ATI Paragon Live Sound consoles (one of three that he owns) parked in the PRS showroom.



It's similar to this later generation console. The two power supplies had been sitting on the floor while I was restoring the vintage amps of Mike Campbell's until I was thru with that project, then I was moving them up onto the bench to dig into them. I had no idea I'd have to pack up and move back to my small shop down the street before I could get to that task. I borrowed a small empty road case for Fender Deluxe Reverbs to set these on to move them. Almost had it tip over, as the left side of the supplies is all power xfmr/chokes, so totally left-heavy.



Winn managed to find a cable set that had a mating male that had two supply cables breaking out of it's housing into a pair of female connectors for $50, so I had the hardware to cobble a power supply break-out interface box. I've never seen this particular connector before (19-pin)...I've not learned (or that yet sticks in my mind) who makes the connectors....KC is what's molded on the connector shells. The dual cable set I was given had 14-conductor 4mm gauge wires in them....lighting cables from what somebody told me. These PPI1 power supplies have nine power supplies: +/- 20VDC @ 20A, +/- 48VDC @ 0.75A, (3) 5VDC @ 6A, 12VDC @ 2.5A and 3.3VDC @ 4A. Winn has the full console service manual book, which is a beautifully produced set of documents, though what it has on the power supply is limited to the power supply regulator board & crowbar circuit on the right rear side wall. Nothing in it contains the power xfmr wiring to the multiple bridge rectifiers that populate the remaining space in the xfmr/choke village in the chassis. I was also told one of the two supplies immediately blows the mains breaker when turned on, and the other has issues yet to be found, other than the console doesn't behave properly if used to power it.

With the supplies on the road case, I wheeled it over to my power cart (power analyzer, load bank, 30A variac), connected the top supply and nudged the variac up from 0VAC. Drew current immediately, so that was the one having fault current. Before seeing the insides, I figured that would be the easiest to find. Don't quote me now. The other one, I was able to turn on, and and began to see tally lights on the front panel come on...all but the 12V light. I powered down, moved the road case next to the check-out bench and was able to lift/slide that dead supply onto the bench, while I managed to keep the road case from rolling over from the weight imbalance!

I took the male connector apart, which was a task of it's own, as the screws heads were partially stripped...Philips/Slotted. I got it apart. It had two of these thick jacketed neoprene cables joined together and most, but not all of the conductors were in parallel. I cut the one cables' wires from the other one at a time until I had finally separated it, and it fell to the floor. I set the remaining male connector/cable aside, and opened up the service documents to see what I had to work with for documentation, and how...now knowing I had a 14-conductor cable and not a 19-conductor one, how I'll address the needs of the break-out box.

PPS1_Power_Supply_Sch&Docs-1.pdf

I have a collection of Amphenol & Cannon MS series connectors and shells, and was hoping there'd be a panel mount flange & strain relief that I could use to mount to a break-out box to support this thick cable. Found what I needed, so then went to Pacific Radio Burbank's website to see what they had on hand in Hammond die-cast project boxes. They had a suitable box in stock, so I drove down the street to their relocated store and came back with a box. Connector wiring, layout and metal-work time.



In the power supply, all of the regulator circuits are stand-alone isolated supplies. No common ground between any of the circuits. All of the grounds are joined at the main buss bar of the console, though some of the circuits....the three 5V and 12V circuits get their grounds joined before heading to the buss bar. The +/- 20V and +/- 48V supplies get to the buss bar separately, as does the 3.3V and the chassis ground of the power supply. So, I joined all of the grounds in this break out box.

As there wasn't any color code of the wiring in the diagrams, I had to ring out the 19-pin cable to get that code, added it to my wiring chart, and also had to add jumpers in the male plug to join grounds that would need to get to the break-out box for lack of 5 wires. I used the LED tally light pattern on the front panel of the supply for my break-out box pattern.

I finally had a test cable to work with for tackling the power supply that turns on. Connected that up and moved forward. I first checked to see if all of the supply voltages were present under no-load conditions. The +/- 48V supplies read +/- 47.7VDC, +/- 20VDC read exact, the 5V supplies read 5.63V, the 12V supply read 14V (initially got 0V reading, but that was a loose wire on the Euro-style bare-wire connector terminals), and the 3.3V supply read 3.7V. Next was to see how everything read under load. I loaded the +/- 48V supplies to 110 ohms for 0.44A, and both read +/- 47.3VDC. The 5V supplies I loaded to 2 ohms, and they read between 5.4V and 5.6V each at that 2.7A load. The 3.3V was rated for 4A, so I loaded it to 1 ohm, and it read 3.5V, so that was good. The 12V I loaded to 6 ohms, and it read 13.9V, so it was fine. The +/- 20V supplies I could load all the way to 20A. I started with 8 ohm loads, got +20V, -19.8V. with 4 ohms, I got +18.8V and -19.7V (5A load). At 2 ohms, I now see a regulation problem, now reading +14.2V and -19.5V. I didn't check it at 1 ohm, as I see I'm going to have to go digging, and learn how to pull the motherboard out...I think. First, I have to remove the bottom panel.





The other power supply which pulls gobs of current, I removed the bottom panel to see how it's mains were wired, as well as getting a view of the bottom of the power supply PCB assembly. All of the buss caps are Snap-in type. I had visions of these loaded with computer graded screw-terminal caps, as the vintage was right for that style construction. When I had first pulled the top cover off, seeing that huge Toroidal power xfmr and five EI core transformers (assumed), labeled +20, -20, 5V-1, 5V-2, & 5V-3, it hadn't dawned on me that those weren't power xfmrs. If they were, the size of the Toroidal xfmr didn't make sense, nor die the vast number of leads exiting the tip of the xfmr. I concluded those were inductors before I saw the print on the front panel...High Performance LC Filter.

In checking for shorts on the Input Voltage terminals of each regulator circuit, all being fed from the multitude of bridge rectifiers on the other side of the wall....I came upon the three wires into the +/- 20V circuit. (-) and COM were a dead short! I followed the wires over thru the wall, seeing the RED Com wire enter the Toroidal xfmr. The other two disappeared into the maze of wiring and chokes, and from the top view, finding my way into finding (I hope) a shorted bridge rectifier, it looked like there were two pairs of bridges in parallel from the source of the +/- 20V regulators. But, this connection was showing a dead short between the winding's CT and the output of the bridge. I removed the three wires, and the short wasn't on the PCB...it's still on the two wires. I have an uneasy feeling about that.

So, I have a power supply regulator that's falling out of regulation, with the high current end of the circuit living in the stamped-out heat sink forest on the front of the PS board, while the error amplifier has to be unsoldered from the PCB from the bottom of the board. It could be the age-old solder joint problems that I spend so much time repairing....would that life could be so simple. Still, I need to order fresh power supply caps for these, as the age would dictate fresh is needed.

[nevetslab]

There are eight bridge rectifiers total. It appears only the +/- 20V supply is wired with two bridges in parallel, though it's mostly hidden from view. The rest are all discretely used, with each regulator using one each. The +/.- 48V uses one, so it's com wire comes from the C/T secondary of that winding. I'm guessing the output of the (+) terminal on the three 5V bridges and the +/- 20V circuit feed their discrete inductor, with the other end of that inductor's wire feeding the (+) terminal of the regulator. Except for the +/- supply, where each output from the bridge feeds an inductor.

Now, finding what looks like a dead short between the Center Tap of the +/- 20V secondary winding and the output of the (-) terminal's inductor, there's still a lot of wire in that shorted circuit. The inductor's DCR, the inductor's lead wires, which look like 10AWG, maybe 12. One or two shorted bridges where the (-) side of the bridge output shorts to the input side....and usually bridges short in all four quadrants to make life interesting. If that was the case, I'd be seeing a short between the (+) input and COM, but I don't. Only the (-) lead to the COM lead. This one has me stumped for the moment.

After carefully zeroing my Fluke 8060A DMM's ohmmeter leads for less than 10mOhm (0.00 ohms under REL reading), the 'dead short' between COM and (-) leads at the Euro Terminal block reads 110mOhms (no resolution here). Reading wires from other terminals to where I can reach their bridge terminal blades, so there's a least a foot of 16AWG wire in place, I've seen readings as low as 0.01 or 0.02 ohms. My Valhalla 4-terminal 4014 Digital Ohmmeter stopped working years ago, but would be handy here, as that was what I used in measuring xfmr leads during temp rise testing, as it gave resolution less than 1 milliiohm when zero'd. After drawing the circuit out, 160 milliohms could be the resultant reading thru the shorted diode cells, the inductor winding, the secondary leads in parallel of that +/- 20V secondary winding to the centertap plus the wire resistance. Maybe it is just all that. Won't know until I dig down to find the bridge (s) and unplug the wires.

I did just order fresh snap-in buss caps for both supplies, as well as the smaller output caps from them. Size restrictions on everything, so followed those restrictions. I also ordered 10 35A/200V bridges. What's installed are Motorola MDA2502 bridges, which look so much nicer than the more generic metal blocks or plastic blocks filled with potting compound over the discrete diode slugs wired to the quick-disconnect blade terminals. I've always been a little suspicious of paralleling diodes or bridges, as it doesn't really seem like you've doubled the current rating. When things fail, the weakest link goes first, followed by the other in parallel. So, in this case, two 25A bridges in parallel isn't really 50A. Or am I mistaken on that?

[Leo_Gnardo]

I've always been a little suspicious of paralleling diodes or bridges, as it doesn't really seem like you've doubled the current rating. When things fail, the weakest link goes first, followed by the other in parallel. So, in this case, two 25A bridges in parallel isn't really 50A. Or am I mistaken on that?

I think you're right. Diodes & rectifiers have small variations in their forward voltage, so in a parallel situation the one with lower voltage will do the lion's share of the conduction. Maybe all of it, depending on the difference between the two. Perhaps ... if the rectifiers were carefully tested under full load conditions and matched, you could get away with paralleling. After decades in service baking inside that PSU, maybe not so much. I needn't tell you parts change as they age. Sounds like a beast of a power supply, I wish you the best of luck.

[Randall]

You are a better man than I.

[Enzo]

If I were to parallel a couple of heavy rectifiers, I think I would want some low value ballast resistors to urge them to cooperate. Just like in a parallel SS output stage.

[Jon Snell]

"I need to order fresh power supply caps for these, as the age would dictate fresh is needed."
If they are good quality capacitors 105° and test within specification, there is no "need' to replace them.
If they are bulging or leaking, that is a need to replace them.
Parallel diodes, if modern design, will share the load without a problem as they should be from the same batch and have near identical specifications.

"I removed the three wires, and the short wasn't on the PCB...it's still on the two wires. I have an uneasy feeling about that."
New mains transformer may be required from your description but ... does the transformer take out the mains fuse ... if it does, replace the transformer and IEC/mains connectors as a bad connection in the mains power will cause transformers to overheat and aquire shorted turns that make matters worse until dead short.

"but that was a loose wire on the Euro-style bare-wire connector terminals".

[nevetslab]

If they are bulging or leaking, that is a need to replace them.
Parallel diodes, if modern design, will share the load without a problem as they should be from the same batch and have near identical specifications.

"I removed the three wires, and the short wasn't on the PCB...it's still on the two wires. I have an uneasy feeling about that."
New mains transformer may be required from your description but ... does the transformer take out the mains fuse ... if it does, replace the transformer and IEC/mains connectors as a bad connection in the mains power will cause transformers to overheat and aquire shorted turns that make matters worse until dead short.

"but that was a loose wire on the Euro-style bare-wire connector terminals".

Several of the buss caps are bulging, three had their plastic top covers come off, though none show the classic top vent split open as will happen when they go catastrophic. All of the buss caps are 85 deg C rated. I see date codes on two of them...8926 and 8935 (26th week and 35th week of 1989, and on the triacs of the crowbar circuit...9003....3rtd week of 1990. So, we're talking about a 30 yr old power supply, or thereabouts.

I've now assessed the placement of the eight bridge rectifiers. The suspect pair of bridges that appear to be wired in parallel are the second set down off of the side wall in the power supply 'village', so somewhat accessible. I'll remove the two upper bridges (3.3V and 12V bridges), and see how well I can grapple with the secondary lead wires feeding these bridges. Once I have them disconnected from the bridge, I'll know if the damage is catastrophic or not. AS a replacement xfmr would be prohibitively expensive to have made (ATI is no longer in business), that would lead to buying a known replacement power supply for these consoles from an outfit back east who both services them as well as has designed/produced replacement power supplies...SMPS based design. No IEC AC Mains connector here. 14/3 Black Vinyl jacket power cord. 20A internal Airpax circuit breaker. I'll know more later today,

[nevetslab]

This morning, having added the wire colors on the eight bridge rectifiers' wiring to my notes, I then removed the upper layer of bridges and tucked the wires aside so I could first take photos of the two bridges used in the +/- 20V power supply circuit. Each leg of the secondary was connected to both AC terminals of each bridge, while the (+) terminals and (-) terminals were wired in parallel. I found after removing the eight push-on terminals from the two bridges, the SHORT circuit I was seeing at the (-) & COM terminals of the INPUT to the +/- 20V regulator was NO LONGER showing a SHORT. And, found just one of the four diode sections of the left bridge was shorted. I removed that bridge to confirm that. What hasn't been confirmed is if there is any damage in way of leakage in the inductor. If there was damage to the Toroidal Xfmr's secondary winding that powers this circuit, it hasn't shown up just yet, though I only turned the varaic up to around 15-20VAC.



I'll have new 35Amp 200V bridge rectifiers here in the next day or two, along with fresh buss caps to load into both power supplies. I plugged this power supply into my Check-out Bench Variac/Power Analyzer, and ran it up a ways, finding NO SHORT/Hi Current being drawn, and began to see Tally Lights on the front panel. As I've unplugged three of the bridges from the system (3.3V, 12V & 5V#1 which were the three upper bridge rectifiers in the way to get to this middle layer, I didn't take the AC Mains all the way up to see what else is (perhaps) wrong. One victory at a time. So, made progress this morning.

[nevetslab]

I have the 'working'' power supply turned over and removed it's bottom cover to have a look at the PCB, wondering if there are solder joint issues to be found. I see a few, though not obvious like I'm used to seeing. While I had removed the top cover, this supply looking older than the other one, and ATI on it is Audio Toys, Inc vs Audio Technology, Inc on the other one. Both are using the same PCB, dated 3/18/89, REV A. I also noticed all of the bridge rectifiers on this 'working unit' to be more like the fully-plastic low-profile parts like I had bought in the recent past thru Mouser: Vishay GBPC3504. All the bridges in the one having a shorted bridge cell are Motorola MDA2502, with a different blade pattern Nicer looking Bridge package. The Buss Caps also look like they've been replaced in the past. My parts aren't due in until tomorrow, so today is clean-up time. Insides of both are filthy, so I need to drive to the other building to borrow the Porter-Cable Air Compressor so I can blow out all the debris that's been sucked into these over the years.

Two of the three PHMS Hex Drive screws that mount the side-wall Crowbar Clamp PCB now have radiused 5/64" hex drive patterns, and haven't been able to get anywhere gripping onto the side of the screws, so I guess I have to cut slots across the pan head to (hopefully) unthread the screws. No doubt victims of Loctite. I'll also unsolder the Error Amplifier for the +20V regulator to have a look at it...that being the regulator that's falling out of regulation. OK, cut the slots, though had to try a couple different bladed screwdrivers to find one whose blade wasn't standard thickness for a 1/4" width. Sure enough....locking compound found on the threads.

[nevetslab]

The Mouser parts order came in for these two ATI PPS1 power supplies. All the values and sizes were good. I tackled the two 35A bridges first, just to get that out of the way. Metal case, (+) and AC marked on the one side, no other markings. Usual form. Just to be sure, rang one out to make sure it was as anticipated, and was. I marked the other terminals just for sake of completeness. Getting the part back into the chassis, half-way down in a space made for 3 yr old children....only they don't have dexterity to deal with such. So, there was fumbling. My bag of #8 flat washers was half-full of those for #10, just to aid in non-centering washers. Got the nut started. The 19-pair connector of the rear wall was, of course, in the way to aid in driving the mounting screw of the bridges being installed. Tightening with the small socket wrench transferred torque to rotate the bridge, so the gremlins were on the bench to make it interesting. Did finally complete that task, but won't power up until I have fresh buss caps in place.

That power supply that works (apart from the regulation issue) was ready to have it's buss caps unsoldered. I had first unmounted the ribbon cables with their connectors plugged into the board that feeds the crowbar board on the side wall, which is in the way of the local 5V buss cap at that side edge, so I had that removed, only finding two additional cables soldered onto the power supply board, so left that and tie-wrapped the assembly to the power supply buss wires, keeping it out of the way for addressing the removal of the buss caps.

It was easy enough to unsolder all of the buss caps. Flipping it over to extract the caps...now seeing all were glued to each other, to the PCB, to adjacent parts to aid in keeping them in place, so I had the usual fun with aged RTV. Wasn't as bad as dealing with Ampeg SVT-CL's. Cleaned up the PCB surface prior to breaking out the fresh buss caps and getting them all ready for installation.

What I found while fighting the individual packing practices of different Mouser workers made the effort tedious. I always love it with inner bags being taped to the inside of the outer bag! Guess they hire anyone for workers. What I found were both normal length snap-in terminals as well as short length terminals. Actually, all but those for the 3.3V supply were the shorter ones. 0.140" above the mtg surface, vs 0.240" above the mtg surface (on the longer leaded part). I applied RTV around the rim as I installed each part. The first one (5V smaller dia cap) didn't want to snap into place, it being up against the wall, so I felt like the deck was stacked against me. I finally got all the parts glued and installed, and carefully tipped the chassis up on it's xfmr side to have access to the solder side. The two 20V buss caps promptly left the playground, and I found those would have to wait while I dealt with those who were cooperating. I did have to modify my normal solder practice by getting a lot of solderer onto my 3/16" chiesl tip and applying that solder to one of the terminals of the cap I was holding firmly against the PCB, enough to hold it while I released my hand & soldering iron and got to the solder side to solder in the other lead solidly, pressing the part down in the process while it cooled. Then re-soldered the holding lead for a good connection. I had to do that same process on the two 20V buss caps, and finally had them all soldered in. A fair amount of loud cursing was accompanying this activity. Some tasks are really meant for an assistant, which I don't have. Sigh.......

I still have the smaller output filter caps to replace, then some re-soldering in the +/-20V regulator circuit that I saw yesterday. And, I also need to apply additional RTV to bond them to the floor, each other and whatever else I can come up with for long-term stability. So, progress....I think.

[nevetslab]

After getting the new buss caps installed along with the output caps, I went to re-install the crowbar board onto the side wall. It wouldn't fit! HUH? I looked at the cap that came out, these having been replaced at some point later, though my height criteria was taken from the other chassis which had the short. Those were 50mm, and the crowbar board was virtually kissing the top of the 5V#3 buss cap. A shorter can was installed in this chassis I just re-capped. Two of the three holes of the crowbar mtg holes don't align. I looked at the other chassis, and they were about 0.200" further up towards the top. So, time to break out the long round files and elongate the mtg holes. And, as this will generate steel filing debris down upon the PCB below, I have to mask off the area to capture the debris.



There was still conductive debris sitting on top of the power supply PCB, so I had to rotate the chassis a couple times to get it upside down, then tipped up more to get the remaining debris off the top flange, where it landed after rotation. Brushed all that off the case, and set back down to complete the cleanup. I didn't see this one coming.

Different bridge rectifiers used in this chassis.



I will swap out the two 20V circuit bridges in this chassis, now that I have fresh 35A bridges on hand.

I still have to lay in RTV on this newly installed set of buss caps, and will do that before I leave for the day.



I checked the caps for Date Codes: 5V ckt: 18,000uF/25V 105 deg C, D/C 2020; 3.3V Ckt 33,000uF/25V 85 Deg C D/C 2003; 12V Ckt 22,000uF/25V 105 deg C, D/C 1914; 48V ckt 5600uF/80V 105 Deg C D/C 1839; 48V 220uF/63V Output Caps 105 Deg C D/C 2104 On the Date Code, 2020 = 20th week of 2020. So, Oldest date codes that Mouser supplied me was 1839...39th week of 2018. I don't know what their policy is on date codes relative to 'fresh stock'. Onward........

[nevetslab]

This morning I swapped out the two bridges for the +/- 20V regulators. One WAS a 35A bridge, while the other was a 25A bridge, and was partially defective...the potting compound had elevated from it's original position, and there was a hole in it above one of the diode slugs. That diode measured much higher than the other three. I installed the two new 35A bridges, and, following the wiring color code at the input to the regulator circuits on the PCB (WHT/RED for +, WHT/BLK for -), I finished that up, then moved the supply on the road case, which with the weight being all on one side, still made the road case want to tip over. Got it stable, put the top and bottom covers back on, then plugged the breakout cable in and the pair of multimeters, plugged them into the +/- 20V supplies, and with the variac set for 0VAC, switched on and began walking the variac up. I only got a few volts up when the mains current began rising, and shut it down. What did I do wrong. Unplugged the load cable box, removed the meters, remove the top cover and looked at the bridge rectifier's wiring. All looked correct. I unplugged the upper three....the 3.3V, 12V and 5V#1, and tried again. Still pulling current. So, I removed the two load cables and AC cables to the 20V bridges, and tried again. Now, no current drawn, other than nominal. I plugged the AC input wires into the two bridges, leaving the load cables off. Tried again, and no excess current drawn. OK.....I'll connect one of the loads....the WHT/RED, which would be the (+) cable to the buss cap. As soon as I began turning the variac up, it's drawing current again. Disconnected the mains, unplugged that load cable from the bridge, and looked with the DMM in Ohms to see what I had. WHT/RED at the supply input to the WHT/RED cable. Open circuit. WHT/BLK at the supply input (-) to this WHT/RED cable at the bridge...continuity! HUH? Tried the WHT/BLK cable (-) at the supply input to the WHT/RED cable at the bridge. continuity. Looked at the two 20V inductors. One wire, going to the power suppy buss cap connection is WHT/RED, while the other wire is WHT/BLK, which would need to connect to the bridge (+) terminal. I looked at the photo of the bridge wiring I took a couple days ago to see what the forward wires was, that was connected to the (+) bridge terminal. WHT/BLK. Sigh...........

I didn't realize they did that in the wiring, as I don't have the chassis wiring diagrams for this. Logic is an organized way of going wrong with confidence! So, I reversed the wire colors accordingly, and tried again. Now, no current draw, and I'm getting proper voltages at the two supply input terminals at the end of those cables. Powered down, then restored the rest of the bridge wiring, powered back up to verify all was still ok, then put the top cover back on and moved the supply back to where I could check it under load, though first, I'd let it idle at AC Input voltage steps, since I did mistakenly miswire the polarity to the two new buss caps. I didn't get very far before the ammeter told me stop!!

Finally got it running at full mains voltage, and all the supplies were reading as they had before at no load. This time, with fresh bridges on the bipolar 20V supplies, I no longer had regulation problems. That must have been from the faulty bridge I found this morning. I loaded the 20V regulators to 20A each, and they held up at 19.9V, -19.7V. At 10A loads, 20.0V & -19.9V, so that seems ok. All the other supplies under load were fine, so I think this supply is back to normal.

I checked that other supply that I had wired as I had initially wired this one, and, sure enough, it would no doubt pull mains current if I brought it up enough. I'll correct that wiring in the morning when I pull the buss caps off of it, and re-cap it. Progress.

[nevetslab]

I corrected the other supply's load cable connections, so that task is done. I plugged it into my variac/power analyzer, switched it on, and ran up the variac enough to see if I had all of the front panel tally lights lit, and that it wasn't drawing fault current. All but the 3.3V tally light was lit. I turned the variac back down, and checked the voltage at the 3.3VDC input circuit. it was -1.7VDC, not +1.7V. Turned off, and looked at that 3V bridge rectifier. I had removed the top three bridges (3.3V, 12V, 5V#1), and when I was thru with the wiring corrections down below on the 20V bridges, I put these back into place. I visually went off of the Motorola P/N & Logo facing out, and hadn't looked to verify that front middle terminal was (+) on all of them, as that was their 'standard'. The 3.3V bridge was (-) on that one. Opps. Rotated the bridge and re-connected, tried again. Now all the tally lights light up, and nominal no/load current being drawn, so I have a good start on this second power supply.

I added rubber feet to the bottom of the now-working/tested power supply. I ran load tests this morning and recorded that data for my records. Winn doesn't have any of his ATI Paragon consoles here at the PRS showroom, so I'll have to go with him for final test/look-see. I didn't try to load all the supplies at once, as I’d need to dig out more cables for that task, but I did load everything to full (or nearly full) rated load current, and the regulation looks good on all the supplies. I’ve no idea what the console pulls, but would be able to get readings on the power analyzer for that overall, when I’m thru.

Winn was there at the PRS showroom, so stopped by to see him, and he came back with me to help move the supplies around so I can re-cap this second supply.

The Audio Supplies are all well regulated…..+/- 20V N/L thru +19.7V/-19.5V @ 20A loads, +/- 47.7V N/L thru +/- 47.2V @ 0.43A, with the rest of the regulators producing greater than the nominal stated voltages.



There's a pair of resistors on the back side of the 20V regulators, and while I haven't stopped to see what the value is and where they're placed in the circuit, I'm guessing these were placed to tweak the supplies to +/- 20V. I'd think that would be done on the small error amplifier daughter boards soldered into this board for each supply. But, as these have been working fine all these years, apart from the recent failures, no need to make a full project out of it.

The three 5V circuits, each rated at 6A read 5.5V avg full load, 12V reads 14V full load, and 3.3V reads 3.5V full load. The 5V circuits run the LED bar graphs on the modules, the 3.3V runs the VU filaments, and the 12V runs the fans and lamps. I looked thru the schematics to see if there were digital logic circuits that would want tighter regulation then the 5.4-5.6V potentials, but only saw LM3914 bar graph circuits to run the various local meters for Input, RMS and Compressor meter functions on the main input channel strips. So, one down, and one left to recap, and verify all the supplies are nominal. Second one is usually easier, once the first one has paved the way thru.

Onward…….

[nevetslab]

Yesterday I removed the 30 year old power supply buss caps and output caps from this second supply, the one that had the shorted 20V bridge rectifier. I replaced those earlier this week when the parts arrived. Glued down these buss caps to the PCB, as I did on the other supply, then added more RTV to bond them to each other, as best I could. I can’t say I’ve mastered the art of applying that stuff where you want it to go!



Last night, I found a chart on the Crowbar Schematic that shows the Actual Voltages these supplies produce, which agrees with what I've been measuring, and what the over-voltage trip thresholds are for each supply voltage. Only the +/- 20V supplies are regulated to hold to that value. The +/- 48V listed at +/- 47.8V, the 12V listed at 13.8V, the 5V’s at 5.7V and the 3.3V at 3.7V. This second supply under 50% load hangs in at +/-20.0V, and at full load (20A), is +/- 19.7V. The +/- 48V at 50% load is +/- 47.5V, the three 5V supplies at full load are 5.5V, the 12V at 50% load is 13.9V, and the 3.3V at full load is 3.5V. All the supplies hit regulation at around 80VAC, so there’s good AC Mains regulation.





Today, I checked the noise level on the Audio Supplies (+/- 20V, +/- 48V) at 50% load. The +/-20V supplies measured 35uV RMS, with a little bit of bridge conduction spikes present. The +/- 48V supplies measured a bit higher, not showing any waveshape, but there was VLF modulation causing the meter to read higher….around 100uV RMS on the +48V supply, 50uV RMS on the -48V supply. So, that looked respectable. I’ll wait to hear the console and Winn’s reaction, as he knows these consoles very well.

Looking at the Crowbar schematic, I see 40A SCR’s placed across the +/- 20V and +/- 48V supply lines, where shorting those would no doubt pull enough current to trip the supply’s circuit breaker (20A Mains Breaker, probably 25A trip…I didn’t check). There’s also a relay on this crowbar board with connector P15 shown as 120VAC. The N.O. terminals of the relay are across that 120VAC connection. I hadn’t noticed ANY heavy wires connected to that circuit board that wander down to the AC Mains power switch wiring underneath the large Toroidal xfmr, though just now, looking at the photo of the crowbar board mounted up on the right side of the chassis, I see a pair of GRY wires that I hadn't noticed. After opening the cover back up, I see those two grey wires connect to a thermal switch (COR) on the heat sink fin next to the middle wall I don't see that T/S on the schematic. P15 is visible on that PCB. Not sure what that portion of the circuit does, though the four 40A SCR’s look like any one of the four audio supplies would trip the four to fire. All of this AC Mains wiring isn't shown in the Service Manual for the console. I'll see if I can come up with additional documentation and details on this supply from one of Winn's contacts back on the East Coast who thoroughly knows these consoles and their power supplies.

Also, on the way in, I stopped off at Home Depot, having found they carry 10/12AWG 1/4” female fully insulated quick-disconnect terminals and bought 20 pcs, so I have enough to replace the ones that show burn damage on the outer insulation of the terminals on that unit whose bridge shorted. It looks like prior damage had occurred, from what I saw installed. I didn’t have any terminals on hand, and Pacific Radio down the street only had 14-16AWG size. So, at least I’ll be able to clean up the cosmetic damage that I left in place, connected and working for the time being.

[nevetslab]

Sunday morning, did my weekly early-AM shopping, witnessed a shopper getting busted by the store manager for not wearing his face mask, though the remedy given the shopper wasn't anything an adult could figure out how to place it onto your face! We're back to that in LA, I see. So, I got back home, put my supplies away and made breakfast, changed batteries on my CEL-550 Hand Held SLM/1/3 Oct RTA to bring in for looking at the fan noise of these supplies, as I recall there was comment about one being noisier than the other one. It got closed up in a closet, and later, it was the one that overheated and failed....the one with the shorted bridge rectifier, I guess. That was the one that, while getting it back up and working, there was burn damage on the insulators of the quick-disconnects, that I used as-is for the moment, not having the replacement connectors. So, I replaced those this morning, after first trying that out with two of the larger wire leads I had clipped off/re-terminated. The replacement YEL 10/12AWG terminals fit into it fine, so corrected that wiring to put that to bed. I moved that YEL/BLK wire to straighten the rectifier terminal after seeing the photo.



Then, after having to replace the eight AA batteries in the SPL meter/RTA analyzer, not having had it powered up for months, I had to go thru some of the set-up chores that try my memory. After setting up the instrument and then one supply at a time, powering them up at idle, I took a number of readings, storing the spectrum on each one. Then, I had to figure out how to recall those spectrums, which isn't obvious. Had to open the book to figure that out. I don't have the other end of the interface set up, so I had to resort to taking photos of the display. Another exercise in trying to minimize overhead light reflections, as well as yours truly in the reflection, so I did at least get something close enough to use. The data is still stored in the meter, so I can always try again later.



I had earlier measured the two with a different handheld SPL meter, and got similar readings at 1 meter, only no frequency spectrum. So, we're looking at the two DC fans....one is moving more air, higher velocity. As luck would have it, the Fan's mfgr label is face-out, and not readable from the rear panel. I'd have to remove the fan to see what it is, and.....as luck would have it, hardware access is a problem with one of the screws. The rear panel can be unmounted, though that leaves wiring to the three bridge rectifiers mounted to the rear panel, along with the power cord, so haven't tempted fate to attempt removing a fan. I'll speak with Winn in the morning to see where we are. There is a 4.5dB WB difference in the higher velocity fan.

It's still too hot outside to go home and brave the heat in the apartment, so I guess I'll hang out and listen to the client rehearsing next door.