Ampeg SVT-7pro power shutdown

[mnow]

Got a SVT-7pro that has one of the smps chips getting hot (IC U2) and going into thermal shutdown.
If the tube is pulled in the preamp, the amp will stay on a bit longer but the IC still gets hot.
Checked the caps and diodes on the supply and all are good.
Transformer T3 gets a bit warm also.
Has anyone else had problems with this switcher lately? Maybe a failing U2 or possible bat tranny, not sure yet.

[Enzo]

Hi
Remember, SMPS are EXTREMELY dangerous to work on, use an isolation transformer if you have one, and remember that the "ground" for the U2 circuits are actually the negative side of the rectified mains


If U2 is getting hot I would replace it.

T3 getting hot MIGHT only be the result of DC through it caused by a shorted U2. On the other hand it has several secondaries, I see five rectifier diodes looking at it, and a shorted diode there could cause trouble or even a load problem on one or more of those secondaries.

[mnow]

U2 is not shorted, it just keeps getting hotter until it goes into thermal.
If the tube is pulled, U2 will run a bit longer before it overheats (I would say the extra load to the +15v rail from the filament is the cause)
Diodes and such are all good, all secondaries are good and clean (as in no dried up caps).
Going to throw a 2A265 in there as it is rated for 32W compared to the 2A165 at 18W. Might have been an oversight on Louds design.

[Enzo]

SHorted is a relative term here. You may not find any of the IC pins dead shorts to each other, but inside that IC is a relatively complex circuit, and all it takes is an internal diode to be shorted, and the thing will no longer work right, but will check fine between its pins. The internals are discussed here on the data sheet.
http://www.mouser.com/ds/2/196/ICE2x...2011-92922.pdf

If the IC gets hot, I;d replace it.

Sure pulling part of the load off might make it run a bit cooler, just as if your car is overheating, and you turn off the A/C or release the parking brake, and the engine might overheat less.

If Loud made a faulty SVT7 design, then they would ALL be overheating. You can try a higher rated part here, but that sort of change can sometimes create its own complications. I think it is way premature to call it a design flaw.

[km6xz]

How did you test the diodes? They need to be very fast switching and if not, due to slight additional capacitance, everything is going to heat, yet test fine with a DC reverse leakage test or even sweeping it with a curve tracer. A shift in oscillator frequency will cause heat and losses in the transformer. If you start changing parts in a switcher before you have it repaired, good luck but no one would probably be able to help find the problem then. If you want to redesign it, do it AFTER you have a baseline sent of measurements of the working unit, after it has been repaired. Random experiments and switch mode supplies do not belong in the same sentence, any shift in parameters can keep it from working or being stable.

Since one of your observations is pulling the tube reduces the rate of heating, a reasonable question would be what is the loading of the supply by the amp, and what is normal? It might not even be a supply problem. Is it pulling more than normal current from the mains? Switchers stay cool as long as there is nothing slowing down the pulse rise time, but if anything does, like transistors or diodes with more capacitance than normal, they can get very hot very fast.

[mnow]

OK, Replaced the ice2a165 and it overheated and shutdown. Replaced it with the ice2a265 and it got hot but did not shutdown (still don't like the hot IC)
The preamp board is only drawing another 70ma on the +15v and 40ma on the -15V lines. The 300V line is drawing 1.07ma
Does not seem like that much more added to the power load to cause the overheating.
Checked the waveform on the output of the IC but did not see any difference upon extra load.
Checked caps with a DCR ESR analyzer 88A (it has never failed me before) and diodes with a Fluke 75 meter.
Added one of those "stick on" heatsinks to the ice2a265 and it is still hot but not shutting down though.

Any other thoughts?

SHorted is a relative term here. You may not find any of the IC pins dead shorts to each other, but inside that IC is a relatively complex circuit, and all it takes is an internal diode to be shorted, and the thing will no longer work right, but will check fine between its pins. The internals are discussed here on the data sheet.
http://www.mouser.com/ds/2/196/ICE2x...2011-92922.pdf

If the IC gets hot, I;d replace it.

Sure pulling part of the load off might make it run a bit cooler, just as if your car is overheating, and you turn off the A/C or release the parking brake, and the engine might overheat less.

If Loud made a faulty SVT7 design, then they would ALL be overheating. You can try a higher rated part here, but that sort of change can sometimes create its own complications. I think it is way premature to call it a design flaw.

[km6xz]

What frequency is the oscillator running at, if lower than normal which can happen with any change in reactance in switching transistors or loading on the secondary. Putting your scope probe on the driver for the switching transistors is not advised, too much capacitance with a normal passive probe. I am not a fan of shotgunning but in switchers, if any problems are corrected in one, changing the ultra fast rectifiers is a good bet.
The PWM controller can get hot if the rate to too slow, there is leakage in the switching transistors or a bypass. There is an integrated switching transistor that drives the external switching that is the only part that generates some heat except for a internal regulator. Have to checked the supply voltage for the chip? The internal regulator could be getting hot if the low voltage supply is too high. Normally it is not needing to source very much current.
Unfortunately, many of the DC and AF suitable test gear is not that useful with fast rise time switchers, a cap that looks fine for ESR at low test frequencies can be bad in fast rise time applications. Not many of use have pulse generators fast enough to analyze the ultra-fast diodes needed. I had all that gear and more back home in California but not here in Russia, a lot, but not that much specialized instruments.
So in switchers, I do a little shotgunning and replace all the diodes if the switching transistor or secondary filters failed.

[Jazz P Bass]

How about taking a few voltage measurements on the following pins of the IC.
Vcc:
FB:
I Sense:
Soft Start:

[mnow]

How about taking a few voltage measurements on the following pins of the IC.
Vcc:
FB:
I Sense:
Soft Start:

Here's some info Vcc Unloaded = 14.33v Loaded = 14.95v
FB Unloaded = 1.62v Loaded = 1.94v
I sense Unloaded = .026v Loaded = .045v
SS Unloaded = 5.6v Loaded = 5.6v

Checked waveform as to Unloaded/Loaded and not much change.
Will try to attach them.
Chip is running at 145kHz

[spiffybits]

Here's some info Vcc Unloaded = 14.33v Loaded = 14.95v
FB Unloaded = 1.62v Loaded = 1.94v
I sense Unloaded = .026v Loaded = .045v
SS Unloaded = 5.6v Loaded = 5.6v

Checked waveform as to Unloaded/Loaded and not much change.
Will try to attach them.
Chip is running at 145kHz

5.6 volt on the SS pin is too damn high. I read me a bed time story called "datasheet" and this sucka's only supposed to go to three damn volts. The chip is shutting down because it's seeing an over current condition during Soft Start. You may have an open resistor (R45) but, the chip's gett'n hot so it ain't that. Your transformer is probably loaded like a fool. Check them resistor ohms for SS and the transformer windings, and if that's good check the resistance on the output rails to ground after the diodes. If you've got low resistance after the output diodes yank off inductors (L4, L5, L7) and the linear regulator (U1) and power that motha up. If it starts after any of those are removed, you've got a bum rail.

Gurd luck.

[leed2711]

Hi 1st time poster just wondering if you got to the bottom of the issues with your svt-7 pro i have one with u2 that looks like its gotten too hot and r46 looks like its had some heat the amp itself just stopped working made a clicking sound

[leed2711]

Hi just wondering if you sorted out your dvt-7 as I have one with exactly the same fault cheers Lee

[km6xz]

If it is making a clicking sound it is trying to start up but high current conditions are being detected so the protection shuts down the supply, until it checks again, and shuts down again. Many things can cause that, shorted power transistors in the output is one likely cause.
As was mentioned earlier in the thread, these are dangerous and any slip of a probe or even probing the wrong part can blow it up, if it does not kill you first.
Do you have a lot of experience with SMPS's and the proper test instruments like an isolation transformer? If ever there was a amp that earned the sticker "No user serviceable parts inside", it is this model. The reason is small switchers can easily tell when an abnormal load is encountered and shuts down. This amp has to supply really high current in normal operation so the chance of a major meltdown due to a fault that pulls high but just a little higher than normal and continues to try to operate with when a fault is present. If it is clicking and never starts at all, the best bet is to look for massive shorts, like the output transistors or secondary filer cap.

[nickb]

OK, Replaced the ice2a165 and it overheated and shutdown. Replaced it with the ice2a265 and it got hot but did not shutdown (still don't like the hot IC)
The preamp board is only drawing another 70ma on the +15v and 40ma on the -15V lines. The 300V line is drawing 1.07ma
Does not seem like that much more added to the power load to cause the overheating.
Checked the waveform on the output of the IC but did not see any difference upon extra load.
Checked caps with a DCR ESR analyser 88A (it has never failed me before) and diodes with a Fluke 75 meter.
Added one of those "stick on" heat sinks to the ice2a265 and it is still hot but not shutting down though.

Any other thoughts?

I know this is an old thread. There was never any kind of conclusion so I wanted to chip in my own experience.

One of these came to me with the same problem. It was obvious that the U2 auxiliary power supply chip had blown in the past and taken quite a few tracks wth it. The tracks had been repaired. The chip had been replaced with the Ampeg recommended upgrade and mounted in a DIL socket and it worked again.

It came to me with the complaint that it would run for a while and then stop. The reason was, as here, U2 was overheating. I measured the case temperature at 120C when it shut down. Like the OP I checked current on all the preamp power rails and additionally the LV supplies to the main amp. Nothing abnormal at all. I opted to replace all the rectifiers and snubber caps on both primary and secondary this being one of the very few occasions where such an approach was justified. Didn't help. I then checked the primary current waveform to see if it was rising very quickly indicating a fractured transformer core. Flyback transformers like this are often gapped anyway so it would have to be a big fracture. All looked OK. Tried a new chip, same problem. I learned from speaking with one of the techs at a service center that whenever they see this revision of the board (Rev C) they automatically swap it out for the latest. Clearly this is evidence that these are problematic.

Clearly this amp has run for a while. Something(s) changed to push it over the edge. Maybe the tube was replaced and extra heater current did it, or perhaps the DIL socket reduced the heat from flowing back to the PCB. The data sheet does specify a few sq cm of copper for a heat sink which this PCB does not have. The losses that cause the chip to heat up are due to switching and conduction and I'm betting the lion's share is switching. These losses depend on the voltage and not the current meaning the chip will be almost as hot at low load as it will for high loads. In other words the few sq cm of copper are required.

I hate defeat but I could not root cause this. I glued a heat sink to U2 and it's run perfectly ever since. The case temp came down to 60C. And that my friends might be the practical solution. I added a little RTV just in case the heat sink glue failed to stop it falling anywhere it should not.

[Khefer Kaos]

Hi! I have a SVT7 pro on my workbench with problems with the U2 IC (ICE2A165).
I replace this with the bigger one (2A265) but still with no power. The only way what I can put the chip to work, Is when I disconect all they wires connectors of the preamp board. I first think the problem was in the preamp section, but I connect external supply and there is no problem. Any suggestions?
In the other hand, I used an 8 pins base to place the U2, is that correct? The part behind the chip was carbonized in some points and I remove this charcoal...