Bgw gtc modifications

[nevetslab]

In my collection of equipment I have a BGW GTC Power Amp, a product I had designed and derived from the TRIAMP product, designed and built for powering Compact Monitor Systems small 3-Way Stage monitors back in 1989. This GTC had, at one point, been used to drive a film projector, using an Autoformer to step up the output voltage to 120VAC, and run it in sync with a video track. It worked for a while, then died. It was set aside for later-man to restore. That was back in the late 1990's. I finally dug it out earlier this year and assessed the damage. One module only lost two output pairs, while the other module lost all seven output pairs including one of the drivers.



I no longer had any of the full-PCB-size Bergquist K-6 Insulator sheets, so I was expecting to have to use greased mica, expecting thermoset to have set in long ago. I was pleasantly surprised that hadn’t occurred, and managed to get the PCB with it’s set of power xstrs up and off the heat sink without damaging the sheet.

The other amp module which only lost two pairs, I had removed the base and emitter resistors of those two, leaving the other 5 pairs in place, and checked to see if it would power up without any issues. It came up fine, bias still fine, very low DC offset, passed signal fine under load.

One of the omissions I made during the layout of the PCB & circuit was NOT using seven pairs of current sense resistors (one per pwr xstr), bussing them to feed the current limiter circuits (pos & neg sides). I had only used one pair of sense resistors. That was something I had wanted to correct, but never got around to it. Until today.

Also, one of the two amp modules that was in this GTC had a Vactec VTL5C4 retrofitted into the True Clip LED driver circuit (LED of Vactec in series with the collector of Q112) and an extra input series resistor (ahead of R115) to the front end diff pair, with the Vactec’s variable resistor tied to the middle of those resistors, other end grounded, so overdriving the amp would activate the Vactec, yielding an Anti-clipping function. So, I added one to the other amp module, prior to installing all the current sense buss resistors for the limiter circuit.



Physically installing the two sense resistor busses wasn’t difficult, though might be a little fragile, since I had to stand up all the resistors off of the PCB around the perimeter of the module, using 20AWG buss wire & Teflon sleeving between each resistor joint.



I’ve made the mod’s on both amp modules now. I’ll have to wait until tomorrow to check the results. The amp that lost all of it’s output stage also lost most of the emitter & base resistors, so all of those are new, freshly installed, along with the new current sense busses, so I’ll have to check that out after I get the first amp module powered back up to verify all is still ok with it.



I haven’t been able to add a layer onto the Adobe Acrobat 5 page and draw in or deposit a VacTec symbol and wire it up in the drawing. And now, CenterStaging has moved all of their administration offices to the new building down the street, so I no longer have scanning facilities here, without having to go there and back to hand-draw the mod’s onto an existing schematic. I’ll have to see what I can do when I get home in the document end of things.

9008-0760 Amp Schematic.pdf

[nevetslab]

Troubleshooting gtc amp module

While handling the two modules with those added current-sense busses mounted around the perimeter, I did manage to break off one of the resistors at the emitter. I was afraid of that, so after replacing that resistor, I cut some small dia heat shrink tubing, slit it open, and slid it over the resistors, and then, as the resistors are next to the power xstr emitter resistor leads, I spot-tied the current sense resistors of the new busses to each emitter resistor, so I had better mechanical support.



Then, finally getting on to powering up the modules, I first verified the one module that WAS working before. Still working, now with the Vactec circuit added. I mounted that module into the chassis and moved on to the other module that showed it still had problems the last time I tried to power it up. That was prior to removing the PCB and replacing all of the power xstrs & drivers, emitter and base resistors. That module is still not coming up. Normally, when slow-starting these on the variac, somewhere around 50-60VAC, you begin seeing equal current flowing thru the driver xstrs’ emitter resistors, which is where I normally monitor the bias. I’m NOT getting that behavior, but instead, there’s NO current flowing thru the drivers, and then at around 60VAC or a bit higher, I suddenly start seeing positive DC voltage on the output buss. This last time it was around 8VDC when I stopped. If I take it higher, it will fire the Output Triac circuit and pull output current.

I’ve already replaced the pre-drivers Q121 & Q120 (MJE 340’s and MJE 350’s), and just replaced the miller caps at that stage C119 & C118. So, now working back up the chain towards the front end, current mirror stage being next. I also need to monitor the front end to see if it's turning on or not, and if so, what's happening in the next stages. I’m half-tempted to move my dual Lambda LQ-534 package over to the bench. That can provide current-limited 0-120VDC @ 0-1.7A dual tracking supplies, instead of having Boulder Dam present with just the variac & power analyzer for sourcing my operating power. I used to have a lot more bench space in the engineering lab for these conveniences years ago.

I did open up another BGW amp from the Compact series, which had a pair of these same power amp modules, and just briefly installed one of those into the GTC mainframe. It came up just fine, so, I could have just stopped there, or modified it as I had just done with this one module that failed years ago. But, what’s the fun in that?

9008-0760 PCB Layout.pdf

[nevetslab]

After monitoring the current thru R131 (which shows the total current thru the current mirror stage) while slowly advancing the variac, monitoring both driver xstr's emitter resistors and the output of the amp, I kept getting no bias current thru the drivers, while the output would eventually start climbing north. I was getting sensible current thru the current mirror, so the front end is turning on, the zeners D103 & D106 setting up ok. On a whim, I removed the mtg screws for all seven of the TO-126 xstrs in the slotted opening. No difference. I then found a cold solder joint on one of the jumper wires to the positive driver (JP101), and after getting that soldered, now I was seeing bias starting to show on Q123, but not on Q122, it's PNP partner in crime.

Hmmmm.......since I lost all of the outputs and drivers, replaced the pre-drivers, the failure episode would have no doubt traumatized the current limiter xstrs. I replaced both with MPS A-06 and A-56 xstrs and new signal diodes 1N4148, just to see if that made any difference. That was the fault path! Now the amp is biasing up normal. Actually running a little hotter than normal. I then re-installed the mtg screws for all the TO-126 xstrs. That lowered it a little, as I was probably letting them get too hot unclamped.

We've always biased these amps at 400mV across the 10 ohm emitter resistors of the driver stage. Now running around 430mV. Not a whole lot higher, but enough to also increase the current thru the current mirror. So, I'll lift the bias setting resistor R136 and see if I have anything closer after re-setting the bias. In production, we had a selection in 1% values from 10 ohms thru 499 ohms, which I don't have.

So, I'm almost there now.

[nevetslab]

Made a minor tweak in the bias, now holding steady. The vactec's in place do a good job as anti-clipping limiters, and it's driving bass bins fine again. Almost everything done.....discovered the HPF switch on the front panel is intermittent, so I'll have to pull the front end PCB out to address that....but after I've ordered the replacement right-angle vertical toggle switch (C & K vertical PCB DPDT). Works fine with it in bypass.

[Leo_Gnardo]

To let you know someone is watching this, nevets. Heck of a project, nothing I would attempt, ever. You are the BGW wiz!

[nevetslab]

So many of us that chase the puzzles that we know SHOULD be working.....like, I 'checked' all the semiconductors and they all measured good, but, then.....as you're powering up, something is going wrong. I had to go back to another working amp module I had in another compact series chassis, so I could see at what AC Mains voltage the front end wakes up, and then you see the current mirror start conducting and the current starts flowing thru the drivers, and noted those. When I went back to the newly restored module, I wasn't seeing ANY signs of the driver stage coming up, and a little higher on the AC mains, the OUTPUT began going away from 0V. That's what made me suspect the current limiter xstrs and diodes. I never tried to measure leakage around those circuits....just swapped them out and got the correct results afterwards. With a heavy sigh of relief. There's always something on the bench to teach you from.

Swapping out the HPF toggle switch will be a little bit of a pain. Thankfully it's not one of the TRIAMP's.....no fast way to extract that motherboard that fills the entire cavity from inside front panel to just ahead of the fans. And the harness interface has a very specific stacking order/orientation. That WAS a really fun project in creating it prior to release to production.

[budney]

Hi nevetslab I have to say your BGW posts are fascinating. Only a few years ago I learnt about BGW and quickly realised how great all their products were. The GTC is a mystery to me though, I can't find much info about it. Some of the amps I have are BGW 85/100/150/250/750/GTA/GTB, but GTC is something I don't have. What's even more intriguing is the 3 way Compact Monitor system amps. Can you explain some more about them? They have active crossovers built in? From one pic it seems like you can change jumper settings, that is for crossover selection? The GTC to me looks like 2 GTA modules, but with a much smaller power supply? Seems odd to me, as the GTB was a smaller GTA, smaller PSU and less power transistors, so the GTC seems like an odd mix of them both.