The multifunction tester checks your transistors at 2 volts. The Peavey amp uses them at 81 volts. No leakage at 2 volts does not predict no leakage at 81 volts.
An Iceo test at rail voltage test your transistors for leakage. I put 24 v supply in series with used transistors from a meltdown amp, series a 47 k resistor series the milliamps scale of the dvm. Collector to emitter on the transistors, no base connection. The ones that leaked significant current, would collapse under 85 volts. These transistors had already passed the diode test of the DVM.
Ability to withstand thump at turnoff depends on the speaker. My cs800s doesn't thump at turnoff. My M-2600 does, but not enough to hurt a SP2(2004). SP2(2004) rated at 500 W continuous. I use them at 1/4 to 70 W in my music room.
The triac "protection" circuit is a cheap crutch to help the speaker and destroy all the output transistors on the channel when one fails. Michael bean nfet disconnect circuit can disconnect the rail caps from the output transistors and maybe save the other 9 output transistors. Speaker nfet disconnect circuit can save speaker without saving the transistors. Either disconnect circuit can also be used to protect speaker during power up and power down, with an additional input from detectors and timers. Cheapest way to get one of these competent disconnect circuits that fits in the box is IMHO buy a blown CS800s or CS800x or even a QSC CX304. The s & qsc don't cost $90 freight because of the switcher supply. Those $10 speaker protection circuits on amazon/ebay/aliexp have hard contact relays rated at ~20 A AC. The meltdown current from a shorted output transistor is DC and will weld the contacts of a $2 relay together instead of breaking the arc.
Last edited:
I would assume you mean something like this ($10 on ebay), I wonder if this could be used and the relays replaced with something beefier.
It's advertised to detect DC and to delay startup connection to speakers by 1 second.
This one has a 30A relay. I suppose an even bigger relay could be substituted in place if this thing shorts... of course that's risking expensive speakers downstream by trusting the stock relay:
The relays on that device are visibly rated 4 A 30 Vdc. Probably competent to protect a 50 W amp with 2200 uf rail capacitors @ +- 30 v. Hard relays that are competent at 80 vdc & 10000 uf cap are huge and cost about $100. Nfet disconnect is more competent but nobody is selling board with them. See protection threads about building your own. Back to back nfets rated 100 A or greater, APV1122 photo relay drivers or equivalent.
Now that would be a better idea. This board already has logic (detect DC to speakers and a startup delay). I may look into converting it to mosfet instead of the mechanical relay.
That 30A relay in my second picture... do you think that'd be enough to open upon a fault? When the stock triac shorts it dims lights (and of course kills OTs in the process). I wouldn't think a brief 4 ohm load (speaker online for a millisecond until DC detect disconnects) would pull quite as much current as the triac in dead short mode. The OTs may even survive this.
I was thinking about it. Say it went to rail voltage (one or the other) P=V^2/R so 80^2/4 = 1600 watts dissipated through the relay.
And per the actual relay contacts rating requirements (current) V=I*R so 80v = I*4ohms so I = 20A if we have a full rail on the speaker positive. A 30 amp relay would provide an extra 50% over (assuming they're honest with their 30 amp rating).
Last edited:
I wonder if this would be a good idea (to save transistors and the triac):
1. Insert a fuse in line with the speaker positive output. There should never be full rail voltage on the speaker terminals so figure maybe 50V (PV shows 40VRMS as the max output at 4 ohms). So V=IR-> 50=I*4 and I = 12 amps. So I'd start there or even a 10 amp fuse with the intent of saving the OTs.
2. Increase the rating of the triac (SAC187 or 2N6346A rated at 200V and 12A) to a bigger triac that won't blow with a failure. Such as BTA30-600CW3G rated at 30A / 600V.
3. Insert a relay for on/off and protection against DC at speaker just before the fuse and on both ground and hot speaker terminals.
This would make troubleshooting a little more complicated. Maybe #3 (relay board) wouldn't be needed if the fuse blows before the upsized triac and the OTs.
1. Insert a fuse in line with the speaker positive output. There should never be full rail voltage on the speaker terminals so figure maybe 50V (PV shows 40VRMS as the max output at 4 ohms). So V=IR-> 50=I*4 and I = 12 amps. So I'd start there or even a 10 amp fuse with the intent of saving the OTs.
2. Increase the rating of the triac (SAC187 or 2N6346A rated at 200V and 12A) to a bigger triac that won't blow with a failure. Such as BTA30-600CW3G rated at 30A / 600V.
3. Insert a relay for on/off and protection against DC at speaker just before the fuse and on both ground and hot speaker terminals.
This would make troubleshooting a little more complicated. Maybe #3 (relay board) wouldn't be needed if the fuse blows before the upsized triac and the OTs.
I can't read the black picture, but probably the 30 A rating on the 2nd picture is a 30 A AC rating. The arc quenches 60 times a second on AC. The arc just transfers contact metal from one to the other with a DC arc, causing them to weld together. Serious contactors we used on the refrigerator test line to stall compressors with 110 VDC (to test the overload breaker) weighed about 6 lb, cost $150 in 1989.
No a higher rated triac doesn't withstand the current. I blew a b600 400 v triac the 2nd time on the PV-1.3k. Melted the 18 ga jumper I put over the melted PCB trace.
Fuse on one rail will cause DC from the other rail as the feedback tries to compensate for the output not being full on one way. I put 30 A fuses on both rails on my PV-1.3k then put nfets from the 10000 uf caps to the OT board to disconnect them when the Peavey DC detect circuit went off. Couldn't quite fit nfets in, thought I might need heatsinks if the cause of the fault was shorted speaker wiring or shorted turn in the speaker coil. In those last two cases, instantaneous current might be 100 A or 1000 A, not 81v/4 ohms.
Peavey uses a serious relay in the CS800x and cs800s that they have tested under DC fault conditions. They charge well for them at the parts counter. Cheaper to buy them IMHO in a blown amp that has them. The blown cs800s or cs800x is a better sounding amp than the PV-1.3k (Or CS800(C)) which is where I ended up. Working CS800s was $120, low wattage due to tired PS caps, new PS caps about $30.
I was getting positive rail voltage intermittently on the left channel's speaker output even with all the output and driver transistors pulled out.
Found a bunch of bad connection points and it works again (for now). I like it that the T-03 transistors can't let smoke out with their in the can design. Too bad they're becoming less and less common.
I had been testing it on an old bad sounding guitar speaker. The voice coil had all the 75v it could handle and is so deformed now that it hardly moves. Maybe the next project will be to recone this speaker. I'm sure a recone kit would be more than the speaker was worth.
I'm thankful for the peavey triac because it saved a fairly good speaker when it went to rail voltage the first time. Great idea Peavey chief engineer Jack Sondermeyer who designed this thing.
Side note, I called Peavey today and they were super helpful in finding parts and even had a few tips.
Found a bunch of bad connection points and it works again (for now). I like it that the T-03 transistors can't let smoke out with their in the can design. Too bad they're becoming less and less common.
I had been testing it on an old bad sounding guitar speaker. The voice coil had all the 75v it could handle and is so deformed now that it hardly moves. Maybe the next project will be to recone this speaker. I'm sure a recone kit would be more than the speaker was worth.
I'm thankful for the peavey triac because it saved a fairly good speaker when it went to rail voltage the first time. Great idea Peavey chief engineer Jack Sondermeyer who designed this thing.
Side note, I called Peavey today and they were super helpful in finding parts and even had a few tips.
Definitely track down all of those intermittents that can cause “DC to the speaker”, and get a speaker relay installed before proceeding further. Soild state, mechanical - your choice depending on how conservative you want to be. On amps this size and smaller, I use ice cube types and they open just fine for what typically goes wrong with these things. You hear horror stories of welded contacts but that just doesn't happen as the primary failure mode on an amp with proper current limiting when driving a loudspeaker.
That may be the root of your problems - I’d completely check over the current limiting circuitry. It is SUPPOSED to limit the current to two amps per output transistor pair under short circuit conditions, and about 6 and a half amps peak at full signal. If you are current hogging somewhere or if the current limiting is not working a short could be dangerous. But normally it isn’t supposed to be - the triac triggering is not supposed to blow output transistors or the triac. Some of the older transistor types that were used in these amps (selected 2N3773’s, RCA1B05) would be run slightly outside their SOA but would usually survive. MJ15024’s are completely within the SOA when driving a short. At least they are supposed to be if the current is being shared properly. Of course a failed output transistor can blow the triac, but they really shouldn’t be failing if everything else is as it should be. Once you switch to a speaker relay instead of the triac you have a better chance of tracking all this down because you won’t be constantly shorting the output out.
That may be the root of your problems - I’d completely check over the current limiting circuitry. It is SUPPOSED to limit the current to two amps per output transistor pair under short circuit conditions, and about 6 and a half amps peak at full signal. If you are current hogging somewhere or if the current limiting is not working a short could be dangerous. But normally it isn’t supposed to be - the triac triggering is not supposed to blow output transistors or the triac. Some of the older transistor types that were used in these amps (selected 2N3773’s, RCA1B05) would be run slightly outside their SOA but would usually survive. MJ15024’s are completely within the SOA when driving a short. At least they are supposed to be if the current is being shared properly. Of course a failed output transistor can blow the triac, but they really shouldn’t be failing if everything else is as it should be. Once you switch to a speaker relay instead of the triac you have a better chance of tracking all this down because you won’t be constantly shorting the output out.