Intermittent problem with Carver TFM-35, need help

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I posted an introduction last night. I've been a long time lurker on these boards and have found them to be a great wealth of knowledge.

I'm having a problem with a Carver TFM-35 amplifier that I've been trying to fix for a friend. This is a bit of a long story, but I think that knowing some of the background might be helpful in diagnosing my problem:

This amp was purchased from someone who said that it "worked when it was placed in storage", but it had not been used for 10 years or so. It was cosmetically new-looking, except for burned-out meter lamps. My friend plugged the amp in when he went to pick it up, turned it on, and verified the relays "clicked". When he got home and hooked it up, it played at very low volumes, but began to distort (both channels) once the volume was increased at all. Once the output began to approach about 1 Watt, it would go into protect. I'm a tinkerer more than a technician, but I'm reasonably competent, at least in my own opinion. I generally don't do work on other people's equipment, but I agreed to take a look at this one. I do have a TFM-35 of my own that I had recently re-capped and I thought that to be an advantage in troubleshooting, since I'd have something to compare to.

When I first looked into the amp, it was obvious that the big 10K mfd filter caps were bulging at the top. Unmistakably so. Also one of the filter caps on the 27.5V rail had leaked electrolyte onto the board and corroded several component leads. Board was discolored from heat in the usual areas around the pre drivers and regulators, but also R481-484 all looked burned (those appear to be on the high V rail commutator circuit). I could see no evidence of anyone ever having worked on this amp before.

I started out checking power supply voltages and found the rail voltages to be normal, except for the obvious capacitor problems. I turned the unit off and was going to discharge the caps, because I know on my own TFM-35, those 10000uF will stay above 85V for a long time. However, none of these held more than a couple of volts for more than a few seconds. Next I checked all the heatsink-mounted transistors for shorts and found none.

I continued checking components throughout the amp, mostly in-circuit where possible. I found a couple of small diodes that were open. All the transistors checked good with a diode test. However, EVERY electrolytic cap on the main board was dead. Dead, as in drastically low capacitance. I've never encountered anything else where every cap tested bad. I double-checked my meter... The caps on the input board and the meter boards tested fine.

So, I consulted with the amp's owner. Since no transistors appeared to be shorted, the amp was cosmetically perfect, and it did come out of protection and make sound, we decided that perhaps the capacitors WERE the problem. Still don't understand why they would have all been bad. The amp had obviously been used a lot, perhaps early in its life, as evidenced by the heat discoloration on the board and the burned-out meter lamps. So we came up with a plan: I had saved all the big power supply filter caps from my own TFM-35 that I had recently re-capped. Since there's only a handful of smaller electrolytics there, I suggested replacing them all, using my "used" caps for the bigger ones, so as not to have unnecessary money invested if it didn't work. So I did that. Also replaced the open diodes and the burned resistors I mentioned earlier.

Success! Powered on, clicked out of protect. Set the bias. Measured offset of ~ 8mV and 12mV. I hooked the amp up to (some cheap) speakers and left it playing music at moderate volume for a few hours. No problems observed. So I consulted with my friend again and ordered new capacitors to replace those that I had recycled. I continued to use the amp for the next week or so and had no issues with it at all. When the new caps arrived, I went back in and replaced those, and also fitted LED's for the meter lighting. Checked bias and offset again and declared this amp project a success, returning it to my friend who traded me a nice receiver with a dead channel as payment.

My friend likes to listen loud. He was using the amp to drive a set of Polk SDA SRS-2. He thought it sounded great for a couple of weeks, and then a problem appeared: He called me one day to say that it sounded "scratchy" on the right channel. He brought the amp to me where I work, and we checked the offset and found between .3 and .4 Volt on the right channel. It would still click out of protect and play, but sounded distorted. The other side was still fine. I took the amp home that evening, and removed the bottom cover to check voltages on the heatsink transistors. A few things immediately jumped out at me as odd, but at that point I was just checking and not making any notes. Something was definitely wrong though, so I grabbed a notepad and went back to re-check and record the values, and.... now they were normal. I checked the offset: ~8mV and 12mV. Hooked up speakers and played music: All fine now. I hate intermittent problems. I know that things like this are normally attributable to bad solder joints or broken traces, so I went over this thing with a strong light and a magnifying glass. I re-soldered a few things that looked suspicious, but really didn't find anything that I felt was bad. So, I put the amp back together and returned it to my friend.

To make the long story shorter, this cycle has continued: It will work fine and then the problem reappears. It seems to occur when the amp is driven to higher outputs. Maybe when it hits the top rail. Usually the problem disappears while I'm probing voltages on the heatsink-mounted transistors, but sometimes it continues long enough for me to get a full set of voltage readings from the right channel. One time the problem seemed to persist: I checked everything I knew to check, got tired of it and turned it off and put it away for a while. When I pulled it back out several weeks later, it still had the offset channel. I started probing components in-circuit around the predrivers, but I had forgotten to discharge those dang 10000 mfd caps! Slipped, and a spark.... Plugged the unit back in to survey the damage and the problem channel (the right side) was back to normal! The left side had a .3V offset now, but it was easily traced to one of the 10 Ohm fusible resistors. I replaced that, and everything was back to normal again, until the next time the amp was driven to higher power.

It's been an incredibly frustrating process. I know that everyone is going to immediately think bad solder joints and broken traces, but I honestly don't think that's it. As much as I've looked, I feel like I would have found them.
I've swapped in the input board from my own TFM-35, and the amp will still misbehave, so I think that eliminates that as being the problem. I've searched extensively on the internet for problems such as this, and I've read a few things where people had mysterious issues that fixed themselves..

I get frustrated with it and put it away, but then I get determined to fix it and pull it back out again. I owe my friend his amp back. Maybe I should just give him mine.. I've read extensively on the Carver site and audiokarma, but I think my best shot at help may be on this board. Does anyone familiar with these amps know of anything that could possibly "latch-up", for lack of a better description? And it takes some sort of electrical "tickle" to get it to come back? The capacitance of touching something with a probe, maybe... Anything like that?? Because that's what it seems like. I have NOT replaced the SA1307/SC3467 transistors that seem to run too hot in these amps. I have pulled them for testing and re-soldered them.. The traces around them are fragile, but everything is intact. I don't want to be doing needless soldering/desoldering though. How can I troubleshoot this? I do have a scope. I also have voltage readings from the heatsink mounted transistors taken while the amp was misbehaving and comparisons to the other channel.

Thank you for reading this long thing, and I appreciate anybody's input!
 
Last edited:
Here are some voltage readings from the transistors on the right channel heatsink. These were taken while the amp was showing about .4V DC at the speaker terminals.

Outputs:
Q218 E -0.414V C 28.8V B -0.414V
Q220 E -0.874V C -28.9V B -0.414V

Drivers:
Q212 E -1.86V C 97.8V B -1.86V
Q214 E -1.84V C -98.3V B -1.86V

1st Commutators:

Q222 E 26.4V C 55.6V B 26.4V
Q224 E -26.7V C -97.3V B -26.3V
Q226 E 26.5V C 55.7V B 28.78V
Q228 E -26.7 C -97.6V B -28.9V

2nd Commutator

Q234 E 55.4V C 97.8V B 55.4V
Q238 E -97.5V C -98.3V B -97.5V

Kinda looks like no bias? I don't know if I've checked it while it was acting up, but it always seems to adjust with no problem.

More later. Thank you for any help!
 
Well sorry to say even tough your post is so long provides no information for an average tech to proceed .plus that TBH i dislike the repair procedure very much ....

Here is my 2 cents

---TFM amplifier is made in Taiwan ...Well designed by Bob don't be sure though that the Taiwanese executed Bob's design with teh best of material Check/replace the relays one of them might have damaged contacts ( quite common ).

---The report on the capacitors being all of them damaged is a pointer that amplifier might was exposed to excessive heat for a long period of time that means that soldering is not to be overlooked .

---measuring the offset over and over again will not help in a distorting amplifier you need a scope and a gen to produce usable results...If not available you are chasing a ghost .

----Generic method to troubleshoot this type of problems is heat and freeze and both of them will point you to a bad soldering or a semi that has a mechanical failure ( actually goes open when operating or under voltage while measure ok when static or out of the circuit ) also very common .

---You wrote a lot of things but how about the obvious ?
*verify all voltage present according to the schematic at the time of failure ( obviously no one here is talking about rail voltage )
*Try to make comparison measure at the time of fault with the working ch.
*verify the obvious: all contacts and fuse holders if any exist ( didn't had the time to check the schematic for that also )
*Make sure that bias doesn't drop or disappear totally at the time of the fault result of a broken loose or dirty bias trimmer
*If problems appeared in the secondary PSU there is a chance that in the 15+15V area you have some IC toasted or partially toasted check and replace if needed at no cost

still not much of this is useful next to a scope that will tell you exactly what is going on or point you to a direction where to look .

Kind regards
Sakis
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.