Ampeg SVT-3 Pro - starts up - then fan cuts out before relay switches - no output.

[yankeerob]

Apologies if I have missed a thread dealing with this problem:

As in the title - the amp switches on, fan starts and then cuts out before the time delay relay is switched... It used to do all that but the fan would stay on - the pre-amp was working fine (into an external power amp) but the power amp was very low output and distorted. I have replaced all the 220Ω/10W and 0.47Ω/5W power resistors (they all looked like they'd seen better days), all the output devices, all the 5600uF/80V and 47uF/450V electrolytics and a handful of the Wima coupling caps that were near any appreciable source of heat.

When I first powered the amp on - the fan ran - the relay switched I plugged a cable in and put a little gain and master vol on and tapped the tip of the cable to see if anything was happening and it sounded pretty healthy. I went to grab a bass and the fan stopped running and there was no output. I have cycled it 4-5 times and now it does what I've described in the title.

Any thoughts??

Many thanks!!

Rob

[Leo_Gnardo]

Could be the fan itself is signaling its need to be replaced. When it just plain quits the amp will run as normal - until the output transistors heat up enough to trigger the overheat AC shutoff switch. And I've seen that happen in SVT-3Pro as well as some other similar amps mostly from Ampeg and Peavey. Swap in a new fan & I'll bet all will run as expected.

[yankeerob]

Hi Leo,

Thanks for your prompt response... so would the output devices heat up that fast to cause a thermal shutdown? We're talking the fan quits just before the relay switches...less than 30 secs...

Many thanks again!

Rob

[Leo_Gnardo]

so would the output devices heat up that fast to cause a thermal shutdown? We're talking the fan quits just before the relay switches...less than 30 secs...

With no signal or being played quietly the amp would probably run all day. Start asking it for more serious wattage and the output transistors & heatsink will definitely heat up enough to trip the hi temp switch.

What I usually see in this type of amp is fan comes on at switchon, runs hi speed for a few seconds then settles down to just ticking over. Then as the power section heats up the fan speed ramps up to match. There's a heat sensor for its drive circuit, very clever indeed, so low volume players won't have to suffer with the sound of the fan whizzing at full tilt constantly. The drawback is, as a fan's bearings wear, it takes more current to get it going and the drive circuit isn't sufficient to make it go. Put in a new fan and all works as it should, hope it works for you too. Last one I had to replace was @ 10 years back so I don't have a ready suggestion where *exactly* to find an appropriate fan with part number and all, but I'll bet Mouser or another good parts distro will have something to work well for you in their catalog and won't cost more than say $15. Possibly LOUD, Ampeg's parent company currently, can help out but I hear things are loused up there lately.

[52 Bill]

And what is happening to the power supply voltages at start up? Is the amp going into protect?

There are more reasons than heat to cause the protect circuit to trip.

[Sowhat]

I have replaced all the 220Ω/10W and 0.47Ω/5W power resistors (they all looked like they'd seen better days), all the output devices, all the 5600uF/80V and 47uF/450V electrolytics and a handful of the Wima coupling caps that were near any appreciable source of heat.

Nice comprehensive shotgunning of that output section, however... did you match those IRFP240's and IRFP9240's? Those amps are pretty powerful and nothing creates a hot spot and high current flow like a mismatched power FET. Sure, you'll never get the output section perfectly matched but the closer the better. Anyway, your problem sounds like a protection trip which means that something is drawing more current than it should. Be quick and deliberate here and have your negative meter lead attached to ground, set it for some dc volts, turn on the amp and quickly check speaker output... got any DC? if so, your output section is messed up. If you got no DC on the output then once again quickly check the sources of all 8 output FET's, you can test those easily with the negative meter lead still attached to ground and then probing the 5 watt 0.47 ohm flameproof source resistors running down the middle of the heatsink, (I think it's the resistor lead closest to the power FET's, you'll figure it out). If they are all over the place or if a few seem to be higher than the others then your FET's are not well matched, you may want to address this.

Oh yeah, I forgot the test conditions so here they are:

No input signal
No speaker load connected
A variac and current monitor supplying the mains power or at least use a lightbulb limiter.

All that's left to do is reset the quiescent current or as others would say reset the bias which is posted in another thread on this forum. It's a fiddley process and requires some equipment to make happen.

[MarkusBass]

My thought is that most probably it was not necessary to replace most of the parts, especially capacitors and output transistors. To replace output transistors you really need a good reason (e.g. shorted transistor).
Before you replace anything, you need to diagnose the amp. Otherwise, you may cause another problems and now you are not sure whether caused new problems, or they existed before.
To check the fan you may check the voltage provided for the fan. As far I remember this is 24VDC fan. If it gets less then e.g. 12V, it won't start. And it is supplied through heaters of both tubes in the power amp. So the first thing I would check is the voltage on the fan. If you check the amp in cold environment, the fact that it cuts out may be normal.
You haven't specified whether the amp goes into the protect mode. If not, I would start from checking ECC82 tube in the power amp and the bias of the amp. If it goes into the protect mode, you need to make at least basic measurements (voltages and output transistors whether they are shorted).

Mark

[yankeerob]

That's all great food for thought and thank you all for contributing.

This particular unit has seen a lot of heat - all of the ceramic sealant on the power resistors was cracked and discolored... hence their replacement.

I shall have another poke around tomorrow to see what I can see... I can't thank you all enough!!!

Rob

[yankeerob]

Last one I had to replace was @ 10 years back so I don't have a ready suggestion where *exactly* to find an appropriate fan with part number and all, but I'll bet Mouser or another good parts distro will have something to work well for you in their catalog and won't cost more than say $15. Possibly LOUD, Ampeg's parent company currently, can help out but I hear things are loused up there lately.

Bit more than that these days - for the exact Comair Rotron ST24A3 fan... Not many people keep them any more and the price has risen. I am - as per previous post - going to have a poke around tomorrow to see what's going on with the suggestions given. If everything seems fairly kosher then the fan is the next step but your point is well taken - the fan was noisy... I blew it out with compressed air and it quieted down but it's 17 yrs old and was caked in dust when it came in...

Further to what I've done so far - the 5600uFs didn't look too bad but the 47uFs definitely looked a bit 'puffy'... 17 yrs in a hot amp are long enough for any electrolytic - especially the cheapies used in these...

Again thanks for all the input!!

[yankeerob]

I've done two things that seem to say a lot...

First I measured the DC on the output and there was no more than 3mV... I did this twice. First the Fluke Model 12 was wavering around as it does in DC autorange mode - switched the amp on and after the relay kicked in it came up to wavering between 2 and 3 mV... Switched the amp off and the meter went to dead on 0.000V. Switched the amp back on and it did the same... so.. as far as I'm concerned - a max of 3mV DC on the output is fairly negligible.

Secondly - after that exercise I measured the DC going to the fan from the off (at the PCB header) and it started at 12V and steadily dropped until the fan stopped spinning... twice. I'm ordering the fan and we shall see what happens. I appreciate all this business about matching the FETS and all that but... I paid a premium price for the FETs from from a no questions asked about returning them supplier... I'm guessing they make sure that what goes out the door isn't crap...

So we shall see. I thought it odd that I definitely had a very strong signal come through when I first switched it on after all that work and then it crapped out.

Wish me luck and - I really appreciate all of your help!!!

[Tom Phillips]

...I measured the DC going to the fan from the off (at the PCB header) and it started at 12V and steadily dropped until the fan stopped spinning... twice. I'm ordering the fan and we shall see what happens...

That seems to indicate a problem (or feature) of the fan drive circuit. The fan itself appears to be working as expected. Correct?

[MarkusBass]

That seems to indicate a problem (or feature) of the fan drive circuit. The fan itself appears to be working as expected. Correct?

Yes, there is nothing wrong with the fan. It is 24V fan and with 12V and dropping down it will stop. There are two simple tests to check the fan:
- connect it to 24VDC external power supply,
- warm up the heatsink of the amp with a hair dryer. I think the fan will start after a moment.
As far as I remember fan in the amp works like you described and there is nothing wrong with it. It will start again when the amp warms up.

I still haven't noticed whether you stated what exactly is wrong with the amp (except from cracked resistors). Does it go into protect mode?

3mV on the output show that there was nothing wrong with the MOSFETs. Have you checked ECC82 and bias?

Mark

[mozwell]

check R65 & R66 are ok in the fan circuit
What is IC1 pin 7 doing as the heatsink warms up
It may be IC1 pin 8 that is causing your problem, and all this is linked in with the speaker relay (controlled by IC1D)
Check the voltages on V1 & V2 pin 4 to 5, to make sure the heaters are all working ok during all of this
Is the +65V voltage steady (cathode of D21), is the voltage at the base of Q17 steady
Check R64, R51 & D23, are these ok, do you get a steady +16V on the cathode of D23, is D23 getting very hot

What a weird circuit !

I also attached a "reliability" mod, (from Mr Fahey), so while you are inside this beast, its a good idea to add this in.

[Sowhat]

I appreciate all this business about matching the FETS and all that but... I paid a premium price for the FETs from from a no questions asked about returning them supplier... I'm guessing they make sure that what goes out the door isn't crap...

Matching is less important than it used to be, however, it's nice to know where things stand sometimes... now you know for sure.

[MarkusBass]

I usually buy 2*10 MOSFETs to get 2*4 matched. This is simply because the manufacture process does not guarantee that the parameters MOSFETs are identical. And I want to have currents balanced between the output transistors - this is safer for the amp. Of course, this is not required. Don't do it if you find it not necessary.

If the main problem with the amp is that there is not output after the relay switches on (about 25 seconds after switching on the amp), the first step should be plugging into the POWER AMP INPUT and checking results. You may also check EFFECTS LOOP RETURN input. Do you get output signal in this case?

Mark