I've blown up with speakers connected 3 times. All my mistake.
The speaker protection from the diyAudio store has worked flawlessly.
Highly recommended.
6L6,
Cooking my F5T v2 with mached toshiba grade B jfets, no casscode. I'm only able to get the bias up to .24v on both channels. This seems on the low side. You told nanchangbob to increase R5 and R6 and start over. So if I switch them out to 2.5K or 2.2K, that would almost double the resistance with the pots at full turns.
Don't have the inputs shorted, should I?
One thing is that both N channels took alot more turns than the P channels to keep the offset at 0vdc. Those two are nearly maxed out. I backed down a bit. Still cooking the amp. The bias is creeping up ever slowly. I'd like some fudge room with the pots.
I've read .35vdc is the target. NP said .4vdc for non fraidy-cats, or is that off limits for a noob.
Cooking my F5T v2 with mached toshiba grade B jfets, no casscode. I'm only able to get the bias up to .24v on both channels. This seems on the low side. You told nanchangbob to increase R5 and R6 and start over. So if I switch them out to 2.5K or 2.2K, that would almost double the resistance with the pots at full turns.
Don't have the inputs shorted, should I?
One thing is that both N channels took alot more turns than the P channels to keep the offset at 0vdc. Those two are nearly maxed out. I backed down a bit. Still cooking the amp. The bias is creeping up ever slowly. I'd like some fudge room with the pots.
I've read .35vdc is the target. NP said .4vdc for non fraidy-cats, or is that off limits for a noob.
Last edited:
TboneAK - You are having a common experience, it's a function of the interactions between your Jfet Idss and Mosfet Vgs, and all you need to do is increase the parallel resistors. 
Also, it's normal that one set of outputs (N or P) need more turns to get to the set bias value, as the Vgs of the N's are very, very rarely the same as the P's. That's why there's a adjustment pot on both phases. This also means that your bias will be slightly different on the N side compared to the P side with zero offset. This is normal.
Inputs shorted just helps things from wiggling around when measuring, it's not necessary
BigE - 0.4 is fine if you have tightly matched mosfet and your diodes don't start to conduct at idle bias. It's not "off limits".
Also, it's normal that one set of outputs (N or P) need more turns to get to the set bias value, as the Vgs of the N's are very, very rarely the same as the P's. That's why there's a adjustment pot on both phases. This also means that your bias will be slightly different on the N side compared to the P side with zero offset. This is normal.
Inputs shorted just helps things from wiggling around when measuring, it's not necessary
BigE - 0.4 is fine if you have tightly matched mosfet and your diodes don't start to conduct at idle bias. It's not "off limits".
6L6, BigE- Thanks!!
I’ll play is safe and set it at .35v. How I read it is basically is, higher bias = higher dissipation, out put, and lower distortion.
I’ll try shorting the inputs this next go around.
Quoting NP.
If you are competent, fearless and also own a fire extinguisher, you can find this point. Just run the amplifier into a reasonably low impedance until it gets good and hot – as hot as you plan to let it get - ever. Then adjust the bias to a point below where the idle current starts to really take off. You should find that this point is around 0.4 volts across the 1 ohm resistors. If you are a fraidy-cat, then just set it at 0.3 volts, and conservatively fuse the AC line.
After getting R5 R6 replaced to allow higher bias, and initial biasing, should I load the out puts with a bank of high watt resistors at 4ohms for a final adjustment as he suggests?
Thanks guys!
I’ll play is safe and set it at .35v. How I read it is basically is, higher bias = higher dissipation, out put, and lower distortion.
I’ll try shorting the inputs this next go around.
Quoting NP.
If you are competent, fearless and also own a fire extinguisher, you can find this point. Just run the amplifier into a reasonably low impedance until it gets good and hot – as hot as you plan to let it get - ever. Then adjust the bias to a point below where the idle current starts to really take off. You should find that this point is around 0.4 volts across the 1 ohm resistors. If you are a fraidy-cat, then just set it at 0.3 volts, and conservatively fuse the AC line.
After getting R5 R6 replaced to allow higher bias, and initial biasing, should I load the out puts with a bank of high watt resistors at 4ohms for a final adjustment as he suggests?
Thanks guys!
Thanks 6L6. Maybe I'm a fraidy cat. Or I just have too many output devices on my heat sinks and can't bias above 0.32....
TboneAK, regardless of the value you choose to use, your heatsinks need to be big enough to dissipate the heat generated by the current flowing through the output devices. That is, the heat gain of the sinks has to be low enough to keep the operating current within reasonable limits. That is always a factor.
Please note that at the same bias voltage, hot MOSfets will conduct more current than cool ones. And their temperature will vary depending on the temperature of the room ( ambient temp ). So, if you bias in the winter in an unheated basement ( like I do ), then summer temperatures may bring operating conditions that your heatsinks cannot handle. The effect is that more current flows, which heats the sinks further, causing more current to flow, which heats the sinks further, and eventually, the mosfets are passing so much current, they blow up.
The rule of thumb is to bias so that the heatsinks are never more than 55 degrees C, year round.
This does not mean if you have big heatsinks you should bias above 0.4 V.
TboneAK, regardless of the value you choose to use, your heatsinks need to be big enough to dissipate the heat generated by the current flowing through the output devices. That is, the heat gain of the sinks has to be low enough to keep the operating current within reasonable limits. That is always a factor.
Please note that at the same bias voltage, hot MOSfets will conduct more current than cool ones. And their temperature will vary depending on the temperature of the room ( ambient temp ). So, if you bias in the winter in an unheated basement ( like I do ), then summer temperatures may bring operating conditions that your heatsinks cannot handle. The effect is that more current flows, which heats the sinks further, causing more current to flow, which heats the sinks further, and eventually, the mosfets are passing so much current, they blow up.
The rule of thumb is to bias so that the heatsinks are never more than 55 degrees C, year round.
This does not mean if you have big heatsinks you should bias above 0.4 V.
BigE, 6L6,
All is going well. Cooking at a shade over .3vdc. Went from 1K to 2k2 on R5 and R6. Much more responsive. It's not to that "I barely touched it" point. I'll go in small steps to .35vdc, and see how it goes.
I'm near the ocean (Gulf of Alaska) and it's very temperate and cool here. Not like the mid west where I grew up. Our house temps range 60-70F all year, no A/C needed. Outside temps rarely go much over 70F in summer, or under 10F in winter. 75F and up, the locals are running out of clothes to take off. Those of us white people give a whole new meaning to "White". It's not pretty, sun glasses are needed. It was sunny and 50F today, shorts weather!
That said, the closer I get it to running away lessons its stability if ambient temps are above my normal, or its put into a deferent amp rack/location. I shall stay on the Fraidy-cat side of .4vdc. I'm using the 5U chassis from the store. It huge.
Thank for all the help!!
All is going well. Cooking at a shade over .3vdc. Went from 1K to 2k2 on R5 and R6. Much more responsive. It's not to that "I barely touched it" point. I'll go in small steps to .35vdc, and see how it goes.
I'm near the ocean (Gulf of Alaska) and it's very temperate and cool here. Not like the mid west where I grew up. Our house temps range 60-70F all year, no A/C needed. Outside temps rarely go much over 70F in summer, or under 10F in winter. 75F and up, the locals are running out of clothes to take off. Those of us white people give a whole new meaning to "White". It's not pretty, sun glasses are needed. It was sunny and 50F today, shorts weather!
That said, the closer I get it to running away lessons its stability if ambient temps are above my normal, or its put into a deferent amp rack/location. I shall stay on the Fraidy-cat side of .4vdc. I'm using the 5U chassis from the store. It huge.
Thank for all the help!!
I've been cooking it tonight. No load on the outputs. .38vdc, and 45C on the heatsinks. I can now see how ambient temp changes can effect the bias. Lift the cover off to make an adjustment, and the bias drops a hundredth or two. I have four DMM's hooked up. One for each off-set, and one each for the P-channels.
The transfo is humming now as well. I'm using the Antek 600VA unit. Normal?
I picked up some ballast resistors to make an 80w, 4.125ohm load bank. Should my next step be to back down the bias to .35, and hook up the load bank? I would think that the heatsink temps should rise, which will should cause the bias to rise as well. How fast does a run-away situation happen?
The transfo is humming now as well. I'm using the Antek 600VA unit. Normal?
I picked up some ballast resistors to make an 80w, 4.125ohm load bank. Should my next step be to back down the bias to .35, and hook up the load bank? I would think that the heatsink temps should rise, which will should cause the bias to rise as well. How fast does a run-away situation happen?
Thanks ZenMod! Yes I did build as per papas sch, but without cascodeing the FE. I’m using the Hifi 2000 400mm deep 5U chassis. Would think we want to stay below that point at which the thermistors save the day. I’ll build the load bank and hook it up and see how it behaves at .38vdc. My ACA runs just under 55c, so I keep thinking the F5T isn’t very hot. Hitting .4vdc won’t be hard, the pots aren’t super touchy yet, and I don’t think the heat will rise considerably with .02vdc more.
BigE, I’m not going to test the limits. I’m just getting feel of how this all works and the amps behavior. All this input from you guys is great!
I will most likely set it to .35 and enjoy it. Next thing is to start designing and building speakers for it to drive. The drivers will be here Friday!
I will most likely set it to .35 and enjoy it. Next thing is to start designing and building speakers for it to drive. The drivers will be here Friday!
I have the load bank on the rt channel. 8, 10 watt, 33ohm resisters in parallel for 3.8ohm measured load. The bias has settled to .394 to .389vdc with the load connected. This heat soaking is a slow process. There’s a bit of oscillation, very slow, about .005vdc over about 2mins. Is that normal? I know it’s pretty small and the offset is holding steady. Heatsink temps are still 45c and less. Ambient room temperature is 17-18c. I’m inclined to leave it be at this point.
TboneAK,
Are the values you've reported for how the bias was set with amp case closed or open?
When you close the case, the internal temp will rise, along with the bias voltage.
How long a time has the amp been continuously on at that bias setting? I find that the heat and bias can continue to rise for 24 hours or more.....when I biased, I would bias for a couple hours on the bench, then play music and continue to watch the bias. It is suggested to revisit the bias after 1 day and 1 week.
Note that as you play music, the diodes will be conducting and contributing to the heat gain in the sinks, which therefore will contribute to more current flow through the outputs....
Are the values you've reported for how the bias was set with amp case closed or open?
When you close the case, the internal temp will rise, along with the bias voltage.
How long a time has the amp been continuously on at that bias setting? I find that the heat and bias can continue to rise for 24 hours or more.....when I biased, I would bias for a couple hours on the bench, then play music and continue to watch the bias. It is suggested to revisit the bias after 1 day and 1 week.
Note that as you play music, the diodes will be conducting and contributing to the heat gain in the sinks, which therefore will contribute to more current flow through the outputs....
BigE thanks for the insight. I was already considering that I should back it down some and let it cook over night. It’s now my plan for this evening. The numbers reported are with the covers on.
I have some new drivers showing up tomorrow that will need braking in. Seems like a good opportunity to let the amp cook while driving them.
I have some new drivers showing up tomorrow that will need braking in. Seems like a good opportunity to let the amp cook while driving them.
Hi all, I have the bias sitting at .38vdc with a heat sink temp in the low 40’s C. I have some 86dB speakers hooked up to it, and a miniDSP 2x4 as a pre and has .9Vrms max output. It’s pretty quiet, pretty sure it’s not playing at 86dB @ 1m. The only thing I knowingly changed for this build was R5 & R6 to 2.2k. Do I need more gain on the input side. I still a noob to this hobby and still have tons too learn.
I am going to build the Mezzmerise or the BA-3, and the Yarra. I was leaning towards the Mezz, but it’s a buffer with low gain output.
I’m not sure if it’s the lower efficiency or DSP.
I am going to build the Mezzmerise or the BA-3, and the Yarra. I was leaning towards the Mezz, but it’s a buffer with low gain output.
I’m not sure if it’s the lower efficiency or DSP.