Service Data Request: Kenwood KA-5700 Integrated Amp

[bob p]

thanks Steve. I've spent the morning reading at DIYAudio and I've found some useful information. This one covers Class H designs. I haven't been able to find a clear, conclusive answer about whether or not the problem is there in the Carver amps. From what others have said, it seems that Bob Carver's implementations improved to the point that the problem was eliminated in some amps. Which ones? I don't know. regarding audibility/scopeability of switching noise, the opinions seem to be split about the Carvers with knowledgeable people on both sides. I guess I'll have to hook my 1.5 up to the scope to know for sure.

there was an interesting post about failure modes in Class G/H amps. evidently, many of them are pretty bad when it comes to working on the top end of a bi/tri-amped system. i guess that the commutator switching is designed to be driven by LF inputs, and if they aren't there the HF inputs aren't effective enough in switching the rails. sometimes large HF transients would lock up the commutators, and then bad things would happen. This is supposedly a problem with some amps more than others. Not sure exactly which ones.

[Alexander]

Hey Bob-

FWIW a great amp to use for breaking in speakers, testing cabs and all around general shop use is a used Gallien Krueger RB800. Very robust build; hell, touring bass players love them! Mine is 13 years old or more, only needed a repair once. Decent jacks & nice heavy duty PT that you WISH your Japanese mid-fi amps had! RB800's on ebay are frequently under 350.00 and lower. Even if output is blown they're wicked cheap to fix. Just a suggestion...

HTH,
Alexander
Retrodyne Amplification

[bob p]

I used to service a lot of Carver, we sold a lot of the PM1.5s. Had one in for service, and I fixed whatever was wrong with it for the guy, and then put it on the burn in bench. I was not all that familiar with the amps at the time. I applied a sine wave - probably 100Hz, my all the time favorite freq - ran it up to 2/3 to 3/4 power and left it there warming my loads. A few minutes later the fan goes a mile a minute as the thing tries to cool itself off. I coulodn't figure out what was wrong. It seemed to meet all the specs and seemed adjusted right. Customer was getting impatient as I kept his amp another day, trying to fathom this thermal problem. I called Carver for advice, and they told me - SINE WAVE??? No it won't put up with a sine wave for very long. Music isn't a sine wave. SIne wave is a lot harder on an amp than music. The amp was working perfectly all along after my initial repair. Live and learn.

I've got a couple of those PM-1.5 amps, one if which is nice and quiet, one of which has a horribly noisy fan. The fan on the one amp has gotten so bad that I thought I should do something about it.

I actually found the proper replacement part. The fan is a small 12-32 VDC fan that's rated to pull about 80 mA.

I have no idea what kind of idle voltage the PM-1.5 fan is actually supposed to see. In looking at 20-year old amps, the power measurements across the fans at idle has been all over the map.

I've got one amp that idles at about 6 to 9 VDC. Although that fan is nice and quiet, the low voltage doesn't run the fan especially well. There's a lot of RPM variation. And the fan doesn't move that much air. A second amp had similar voltages, but it varied from very quiet to horribly squeaky/grinding.

I installed the replacement fan in the noisy amp. Now the idle voltages is LOTS higher -- about 18 VDC, and the fan is LOUD! I mean, like an airplane taking off. It moves lots of air, but its way too noisy for anything but stage use.

for home use, I've tried dropping the input voltage to the fan board with a power resistor, replacing the 1-ohm input resistor (R1 on the fan board) with a 10-ohm power resistor. That brings the voltage down from 18 to 14 VDC and tames the ambient noise considerably, yet gives the fan enough speed to run very steady while moving a considerable amount of air.. I tested the amp under load, and the fan RPM seem to increase appropriately. As a safeguard against such voltage manipulation, there's a thermal switch on the chassis that will bypass the dropping resistors and run the fan at full speed if the amp becomes too warm.

Do you have any idea what the proper spec is for the supply voltages to the fan assembly? My oberved measurements have been highly variable -- I've measured values that are all over the map. Most of the voltages with the old fans seem to be too low, and the new values are much higher. I'm just trying to determine what the target operating point is supposed to be.

Looking at the schem, the fan board is supplied by a full wave rectifier off of the low voltage rail on the main PSU card. Rectified AC then leaves the PSU board at point "C", and is passed to the fan board via a red wire. At the fan board, the rectified AC goes through some switchable power resistors to drop the voltage, and then to a single 470uF 35V cap input filter before going to the fan. at least I know that the DC is supposed to be somewhere below 35 VDC.

No answers from CarverPro yet...

any ideas?

thanks.

[Enzo]

Offhand, no. Does that amp have the fan speed pushbutton on the rear? If I recall right, there was the normal low speed trending to high as power increases, or just high speed constantly - switch selectable. Didn't push the button in, did you?

I have no idea the voltage to expect. I'd deal with it if a fan came up lame.

Looking at the block diagram, it seems the fan works off a dedicated 50VDC rail from the main power supply. Look on the PS drawing, right off the PT a couple diodes sorta under the first bridge, making rail "C". Nominally 50V. Back to the block diag. The drawing shows a 1.3 ohm 7W, and then a pair of parallel 47 ohm 2W in series with the motor. A thermal switch will short out the parallel pair over 50 degrees centigrade. From there, motor to ground, with 470uf/50v cap to ground.

SO as I see it, that "50v" rail will vary as the PT is pulsed. No idea what the up and down limits of that might be.

There is a small additional schematic of the fan power board. it shows the 7w as 1 ohm, the two parallels are now 24 ohm each 2W. And the 470uf is now 35v. And between the parallel pair and the 1 ohm is added an 18 ohm 2W with a switch that shorts it out. That is the fan speed switch.

If you don't have these parts of the drawing set, email me at
tmenzo at msn dot com

If I had to figure it, I'd measure the motor DC resistance, and then monitor the actual voltage at the 50v rail, then using motor resistance and the resistors in the little circuit, calculate the voltage drops.

[bob p]

Yes, I've got the drawing you've referred to. I've attached it below to make things easy for us.

Yes, I see that the rail is marked as 50 VDC, but in reality its not anywhere near that at idle. In one amp (new fan motor) its 20 VDC at idle, and on another amp (old fan motor) its sometimes 6 VDC or 9 VDC. Its interesting that fan supply voltage goes up when the amp works under load.

Yes, I've got the push switch on the back. OUT is low speed, IN is high speed. Both positions are variable speed, in that increasing current through the PT will speed up the fan in both settings. If I jumper the normally open thermal switch, the fan will go to full speed (which is much faster and louder) regardless of the switch position. That makes sense in light of the schematic.

The DCR measurement for the motors are:

Original: 10.5R
Replacements: 47R, 53R, 64R; Average: 55R

As you can see the Z value for the new motors is much higher than the original.

I tried adding some Z in series with R1 to bring the fan speed under control. Here's what I got:

Z    Volts   comment
1R    19.8   Sounds like a Lockheed Constellation taking off in the living room
10R   14.6   Still too loud, but quite an improvement
72R    8.1   Very quiet, fan speed seems to be unreliable

I took a measurement with 15 to 25R subbed for R1, but I seem not to have written down the resulting voltage. I'm guessing it was near 12 VDC.

I'm thinking if I could get the idle speed down to 12V that would be a good compromise between adequate fan speed/cooling and noise levels. I'm guessing about 25R of additional series Z might be the ticket.

Looking at the schematic, its probably not a good idea to change R1, as that would hamper the speed of the fan when the thermal switch closes. To modify the low speed setting, I think that the thing to do would be to increase the value of R2. I'm guessing about adding 25R to increase R2 from 18R to about 43R. That way, the fan would be quiet in the slow position, but if the thermal switch kicked in, it would be free to run at full speed until the chassis cooled down. I honestly don't know why R1 is even there, unless it was just necessary to provide some minor adjustment to tune the speed of the original motor.

I'm not sure how well I can math my way to the answer on this one, as the voltage supply rail moves all over the place if either: a) the amp is under load or b) the internal Z of the fan is changed by a fan swap. The supply rail is so reactive that even changing the fan motors was enough to cause a shift in the supply rail voltage. Maybe the best solution is to just put a rheostat in place of R2 and take some emperical voltage measurements while varying the Z of the rheostat with the new fan in circuit.

[Enzo]

Just guessing, but I'd say R1 likely does a couple things - decoupes the fan from the main PS some, and acts as a current limiter for stalled motor??

I have that drawing, but looking at the original amp drawing on the PS page, I note the parallel resistors - R3,4 here - were called out 47 ohms. So there was twice the resistanc ethen as now. That would slow the fan down, eh? The original had no speed switch.

I expect that 50v rail to bounce around, and to only be 50v at full amp output. Remember this is not a conventional power transformer circuit, it is the magnetic resonance whatzit, and the current through the primary of the thing mirrors the draw on the secondaries. So the harder the amp cranks, the more the PT runs and that 50v will climb. But that is what we want anyway - the harder the amp runs, the more fan coling we need.

Maybe you already said, but are your resistors the same value as the schematic calls out? It might be that the reason for difernt resistor values in different schematics is to accommodate different motors use dover the production run. MAybe?

Later PM amps had two fans, for better cooling and then niether one had to run at takeoff speed.

Yes, many folks have asked how to quiet the fans in those over the years.

I agree that changing the resistance in series would be the method, but don't ask me what values.

[Steve Conner]

1R 19.8 Sounds like a Lockheed Constellation taking off in the living room

Looks to me like the replacement fans have higher voltage, lower current motors than the originals, so the dropper resistors have much less effect. However, remember that the behaviour of a motor is affected by the back EMF constant (volts/RPM) as much as the DC resistance, if not more. But a higher DC resistance may well mean more turns and hence more V/rpm.

I have a hate-hate relationship with cooling fans in stereo amps. I built one with a cooling fan, and now wish I hadn't. It's an AC fan, and I managed to get it running slower and quieter by putting capacitors in series with it.

My Ninja Toaster guitar amp also has a small squirrel cage blower in it that makes the most annoying sound ever. I slowed it down with series resistors, but now it overheats when it's in the flight case.

[bob p]

Yes, many folks have asked how to quiet the fans in those over the years.

I've never had a problem with the noise level caused by the original fan. That is, until it became defective and started squealing. So I'm not really trying to change the noise level of the normal fan, which I think is appropriate. What I am trying to do is to scale the voltages of a replacement fan so that its noise level is comparable to the noise level of the original fan. So I don't think this falls into the same category as those people who have complained that the amp had noisy fans.

I'll check those diagrams and follow with more info.

[bob p]

I have that drawing, but looking at the original amp drawing on the PS page, I note the parallel resistors - R3,4 here - were called out 47 ohms. So there was twice the resistanc ethen as now. That would slow the fan down, eh? The original had no speed switch.
...
Maybe you already said, but are your resistors the same value as the schematic calls out? It might be that the reason for difernt resistor values in different schematics is to accommodate different motors use dover the production run. MAybe?

Enzo, you make a good point about the discrepancies between the PM-1.5 power supply schematic and the PM-1.5 fan board schematic/layout diagram. There are some differences in the resistance values. To address those discrepancies (and to provide reference material if anyone else should need it later on) I've copied that section of the power supply schematic and attached it to this post for reference. See below.

So here is how the resistor measurements compare:

[b]Resistor      PSU Schem      Fan PCB Schem     Amp 104xx    Amp 68xx[/b]
R1            1.3R 7W        1.0R 7W           1.0R 7W      1.0R 7W
R2            23.5R 4W       18R 2W            18R 2W       18R 2W
R3            N/A            12R 4W            12R 4W       12R 4W
C1 (filter)   470uF/50V      470uF/35V         470uF/35V    470uF/35V
C2 (motor)   .01uF/100V      ---               .047uF       .047uF
C3 (earth)   .01uF/100V      .047uF            .047uF       .047uF

Notice the following differences:

1. The main PSU schematic (03-06-83 "120 VAC preliminary") doesn't have the speed switch, or the resistor R3 that forms the voltage divider with R2. R2 and R3 set the fan speeds at the two switch settings.

2. The Fan Board schematic (12-08-83) has the speed switch. It also adds R3, which is used with R2 to set the speed settings for the fan. It doesn't show the ceramic motor start cap.

3. Amp 68xx and 104xx actually have two different fan PCBs. Although the layouts are different, they both appear to trace the same.

4. The actual amps have the ceramic motor start cap that is shown in the original drawing but is not shown in the later drawing. Its 0.047 instead of 0.01 as shown in the original diagram.

5. When installing the new motor I put a 0.01uF ceramic cap across the motor terminals. Its what I had in stock.

[bob p]

Maybe you already said, but are your resistors the same value as the schematic calls out? It might be that the reason for difernt resistor values in different schematics is to accommodate different motors use dover the production run. MAybe?

I'm guessing that the fan circuit was modified from the original, because the original/preliminary circuit had a loud fan. Why else add the selector switch that offers the low speed option? Although the modification could easily be explained by different fan motors, it could just as easily be explained by noise. I'd guess that they tweaked the resitors and created the voltage divider to establish a quieter fan because of user feedback in more "demanding" environments. But I'm just guessing on that.

[bob p]

1R 19.8 Sounds like a Lockheed Constellation taking off in the living room

I didn't know how many people would remember the Connie, so I'm glad that somebody got a laugh out of the joke. That was the last of the great Transatlantic aircraft of the turboprop era. I remember seeing one when I was a kid at an airshow at the Gary, Indiana airport. The plane actually belonged to the Black Panthers.

FYI here's a photo that compares the old fan to the new fan.

[bob p]

Looks to me like the replacement fans have higher voltage, lower current motors than the originals, so the dropper resistors have much less effect. However, remember that the behaviour of a motor is affected by the back EMF constant (volts/RPM) as much as the DC resistance, if not more. But a higher DC resistance may well mean more turns and hence more V/rpm.

Steve, the guys at Carver service told me that they no longer stocked the part, but after I bugged them enough, they looked up the OEM name and model number for the original part. Then I had to do a lot of digging to source the part, and of course I couldn't find them in single unit quantities. I had to buy a bunch of them, and now I have a lifetime supply of replacements. Needless to say, I have a vested interest in tweaking the fan circuit so that it performs according to the original specs.

The old part was made in Japan and the new part is made in China. Even though they both share the same part number, the original fan was a single shaft design, and the current product is a dual shaft design. From a physical perspective, both fans appear to be identical on the outside. The dimensions are the same, the tapped mounting holes are the same, and even the oval vent holes on the sides of the case are the same. The vent hole position is very important -- the fan mounting pedestal that Carver fabricated for the PM series amps has identical ventilation holes that line up with those in the fan body. This assures adequate cooling of the fan motor even under low speed under operating conditions. Externally, the new fan is an exact drop-in replacement for the original Carver fan. The installation is so simple that you can even retain the original wiring harness without modifications.

The only tweak, then, is to modify the fan circuit to get the new fan's RPM down to the target level. I'm confident that this can be done by increasing the value of R3 and/or R4. That would be a very simple mod, as it would only amount to a resistor swap on the fan PCB. What surprised me about this was how much series Z had to be added to get the new fan's RPM down into the original fan's RPM range. As you mentioned, the dropper resistors have much less effect.

Do you have any recommendations for RPM reduction other than using larger values for R3/R4?

[bob p]

My Ninja Toaster guitar amp also has a small squirrel cage blower in it that makes the most annoying sound ever. I slowed it down with series resistors, but now it overheats when it's in the flight case.

Maybe you could use an idea from the Carver fan circuit: Mount a normally open temp sensitive switch on the amp chassis, and wire it across your voltage dropping resistors. when the temp goes up the switch closes, bypassing the dropping resistors to increase the fan speed until the chassis cools down enough to re-open the switch.

Just to complete the photo series, here's another photo of the new fan installed in the amp. The second shaft sticks out of the back, but there's plenty of room for clearance. Its quite a nice replacement, actually. The dropping resistors R3 and R4 are the grey resistors located on the board to the left. The 7W cement resistor is R1.

[bob p]

looking at these numbers again, I had an interesting observation about where those resistance values in the fan circuit seem to have come from:

I have that drawing, but looking at the original amp drawing on the PS page, I note the parallel resistors - R3,4 here - were called out 47 ohms. So there was twice the resistanc ethen as now. That would slow the fan down, eh? The original had no speed switch.

Interestingly, the original diagram showed the value of the resistors as 47R, but it used two of them piggybacked in parallel to produce 24R net resistance. Add that to the 1R resistance of the first dropping resistor, and the total series resistance for the fan supply was originally 25R.

Compare that to the newer circuit that uses the switch. In the newer drawing 18R was added in series as R4, and that previous set of two piggybacked 47R resistors was replaced with 4 piggybacked 47R resistors to produce a net load of 12R as R3. The total series resistance of the dropping resistor array is now:

Low speed: R1+R3+R4 = 1+12+18 = 31R.
High speed: R1+R4 = 1+18 = 19R.

In designing the switched circuit it looks like they split the difference, placing values of 19R and 31R around the original value of 25R. (!)

It also looks like they used parts that were already on-hand to do it.

[Enzo]

I remember the constellations. I think President Eisenhower's Airforce 1 was a connie. And in the late 40s early 1950s ther was a popular entertainer named Arthur Godfrey. Sort of a Merv Griffin/Johnny Carson of his day. He was a pilot, and he owned and flew his own connie. He had a variety show on TV, and I recall once he filmed from the cockpit as he flew the thing. What I recall was him shutting down three of the engines to show off how it could maintain flight on one engine.

Nice flowing profile on the connie.

I remember too much of all that stuff, I even recall the Korean war and Adelai Stevenson.