Yes ZM, we did read your post. You described the solution to an "output offset stubbornly stays negative or positive" situaton, but ElFishi's problem was shifting output DC offset on mains variation due to Vgs difference of N and P output MOSFET.
Agreed. However, with all due respect to the design of the M2,
Of course the "some mv" drift should not be enough reason to worry about, I simply offer an alternative for those seeking more fun from diy, no disrespect intended.
Green LED works.
It drops about 1.8V. DC travel now to +-2mV, still monitoring.
I have readings of the total supply voltage from 44.9 to 45.4 V.
Simple multi-meter stuck into wall plug waves from 220 - 228V AC, I have little trust in that measaurement, though.
Math is in line, too:
Optocoupler keeps the two gates 10.5V apart, for zero DC the remaining 34.5V are shared 17.7 and 15.8V by R6+P1 and R7, respectively, as Vgs(P) is 1.9V more than Vgs(N), in absolute terms.
That is, R6+P1 are 112% of R7.
12% of supply voltage variation will thus go to DC output. This is the order of magnitude I experience.
Edit: Voltage across R13 and R14 is also part of the 17.7 and 15.8V.
So 17.7 is dropped over R6, P1, and R13, 15.8 over R7 and R14.
Argument still holds
It drops about 1.8V. DC travel now to +-2mV, still monitoring.
I have readings of the total supply voltage from 44.9 to 45.4 V.
Simple multi-meter stuck into wall plug waves from 220 - 228V AC, I have little trust in that measaurement, though.
Math is in line, too:
Optocoupler keeps the two gates 10.5V apart, for zero DC the remaining 34.5V are shared 17.7 and 15.8V by R6+P1 and R7, respectively, as Vgs(P) is 1.9V more than Vgs(N), in absolute terms.
That is, R6+P1 are 112% of R7.
12% of supply voltage variation will thus go to DC output. This is the order of magnitude I experience.
Edit: Voltage across R13 and R14 is also part of the 17.7 and 15.8V.
So 17.7 is dropped over R6, P1, and R13, 15.8 over R7 and R14.
Argument still holds
Last edited:
Good for you. Just remember to remove the LED when you switch to the Toshibas.
Thx for your idea! I still think about regs, but have to get the parts first. If I stick with LED, it will go to the plug so stay with the MOSFET
Turning off the amp sends ca 2V to the output, does anyone else have that?
I measured all my connections again, and once more cleaned out the space between the soldering pads. I changed the PS cabling from clips to 6.3mm connectors. Looks like the attached now. (removed the other clips I use for measurement, too)
Not sure the 2V would give a thump as it needs 5s to build up and much longer to decay again. Do you know the Vgs of your MOSFETs?
Attachments
I believe you're having a thump without dummy load ?
I don't have any , loaded with slightly above 4 ohms , ~94db/W speakers
dunno Vgs of mosfets , didn't care
whatever you use , it'll have approx. same or even lower Vgs (N vs. P ) difference as with IRFP
OK, hooked up an old speaker. Small thump on turn on, no thump on turn off.
Output voltage only rises to 1V with speaker as load. But as I suspected: too slow to thump.
According to spec IRFP have Vgs of ca 4.9V and 5.2V @ 1.3A, i.e. 0.3V apart. My FQAs are 1.9V apart.
I can not say what will happen in your location, but here in Jakarta where temperature, humidity and pollution is high, regular type of connector will develop oxidation within weeks of use. I avoid connectors whenever possible. Here, silver, gold or rhodium plated connector (which I can only afford a few of) is mandatory to minimize bad contact related problems. Most of the time I solder before crimping the connector and still suffer more than enough share of problems. You have been warned.
Thanks for the warning. These type of connectors seem ubiquitous though in moderate latitudes. Once I'm done tinkering I may consider soldering up those leads. So far it is too much in and out still.
May I return the favor? Where I come from solder under a pressed (crimp, screw terminal, etc) connection is considered bad since the solder will flow under pressure.
I meant soldering the exposed wire to the connector piece before crimping and crimp the soldered joint lightly, then crimp the plastic covered part properly. I can also crimp first and solder the exposed wire to the connector afterwards, but I found the previous method neater and gave a larger contact area. The solder helps to reduce problems arising from oxidation.
That is so misleading and unjustified.OK I'll be careful not to build any amps in Jakarta.
You can always build the amps with all the el cheapo connectors in Jakarta. Then bring it back to your low temperature, humidity and pollution home, office or whatever to enjoy.
Fired up the Toshibas (no LED). Way better. Bought them matched for Vgs (almost) and Gm
Warming up still, but stays within +-1.5mV, thats beyond what I trust my DVM.
Interestingly they run at slightly higher bias, 1298mV across R13+R14, that is 1.38A. Was 1256mV or 1.34A for the FQAs.
Turn off DC excursion is 0.15V only.
In situ I have Vgs = 2.58V for N and -2.11 for P, but I knew that, since NicMac sells them measured. For R7 I measure 46.9kOhm (put in a 47K), for R6+P1 I have 45.7kOhm. For those out there with Toshiba MOSFETs beware though, your values may be different. It just so happens that with my Toshibas the pot would need to go to the other side if you were to stick to 47k for R6.
So having had a PM about it, I might gauge interest in a GB for the Edcor transformers for folks in Australia. As the goal is to split the postage cost from Edcor to oz, I can't estimate what the potential savings would be after repacking and resending within oz. That said it'll be cheaper whatever happens
Please express interest Here
Please express interest Here
Yes, 300VA is fine. You can use the same PSU as other First Watt amplifiers, like the F4 for example.
http://www.firstwatt.com/pdf/prod_f4_man.pdf
http://www.firstwatt.com/pdf/prod_f4_man.pdf