Babelfish M25, SissySIT - general building tips and tricks

Made an interesting discovery today - offset on one channel was in excess of 5v! Pulled the lid to adjust P@ for offset and can't get offset to go below 1.2v. Other channel is great. Luckily, caps in speaker crossover protected the drivers from this voltage.

Any suggestions on where to start looking for problems? When originally set up and dialed in, offset was adjusted to 100mV or so.
 
ref to schm in post #1, understanding that DC Offset is positive, either increase resistor in series with DC Offset trimpot (so R18) or decrease opposite one (so R19)

for R18, you can try 51K


for R19, either change it to 30K or solder 180K across existing one

reason why DC Offset changed with time ......... most likely change in SIT gate current with time

at least that's logical for me, with no clue does or how semiconductor can change behavior with time

fact is, SITs are much more peculiar Animals than all other semis
 
Thanks, Zen! I did go back to re-read post #1 and I was thinking the solution was to increase value of R18 (presently 43k), but I didn't want to jump to this solution, only to later learn it was indicative of a different problem and that I was applying the wrong fix.

I didn't know that the SITs would drift over time. Amp has been used 15+ hrs/day since August when I completed the amp.

Time to warm up the soldering iron again...
 
@Eric ..... well, I can't imagine anything else changing .... all other parts are more or less known, but that elusive SIT gate is showing as finicky .......

you already know that we all practically learned things on the fly, and - as I said several times, it's just wishful thinking - why I didn't get R.3 right away :rofl:

what I learned is - for best results, it's clever to have that gate connected to hefty current path ....... so whatever and whenever gate goes to finicky mood, there is a firm reference to keep it well behaved
 
HA- that's the fun of learning as you go 😉

I might need to do some more work. After powering up and adjusting offset back to zero on a cold amp, it's been creeping up again over the course of an hour as the amp warms. In order to keep offset close to zero, I've maxed out the pot again after changing R18 from 43k to 52k. Maybe it's time to swap out the SIT with another that I have and see if that one behaves any better. Have to let it sit for another hour or two and see where it goes. Then I'll measure again in the morning when it is cold and see what its behavior is.
 
everything counts only in temp. equilibrium

with that type of amp , I'm not much in worry even if seeing 1V5 in cold state, as long it is stable close to 0 in temp. equilibrium

if you have too much problems ....... only explanation is that you have not perfectly behaving SIT ( gate current again)

one of (luckily) rare occasions where I was main offender, made you all pay a price for our collective naivety

OK, reality - great deal of that naivety was all mine


try going the other way - decrease value of R19, increasing current through reference string, in hop it will give more firmer leverage


though, as you said - you have caps i nseries with drivers, so you can be slightly more relaxed with criteria what's proper DC Offset

note - recheck Iq each time you change DC Offset - just in case
 
Thanks, I'll poke on R19 next. Just started a new warm up cycle and making notes on offset and bias measurements at timed intervals to see what happens with bias and offset over time. The goal is to see what the readings are after reaching temp equilibrium, but I'm also interested in seeing the behavior along the way. If R19 doesn't do it, I'll sub in a new SIT.

I didn't realize there was an issue until I hooked up a new set of speakers yesterday and the driver cone popped forward and stayed there 😳
 
Here is the Bias and Offset behavior over a twelve hour period. It seems rather consistent, because I last adjusted the pots at the 2hr mark, assuming that things would settle in after that in terms of thermal behavior - I never measured it beyond the 2hr mark before. Apparently, things keep shifting even after the amp "warms up." With bias this high in the left channel, the Mosfet has been running at 53-55w by the end of the day (right channel wasn't that far behind...). I have 4 more SITs to try, I'll see if any of them have a more narrow swing over time.

SissySIT Bias Offset Behavior.jpg
 
Last edited:
Where the original SIT in the left channel changed in DC offset by 6.5v and bias changed by 215mV over a 12hr period, the new SIT that I swapped in changed in DC offset by only 0.159v and bias changed by only 52mV over an 8hr period today. The new SIT is way more thermally stable and I’m super happy with this change! I need to move R18 back to 43k now as I’m at the opposite end of the adjustment range with the offset pot.

I’ve got one more SIT to play with, I thought I had two extra pairs, but one of the pairs are 2SK182ES devices.
 
Swapping out both SITs did the trick! After adjusting offset and bias, both stay very consistent now. Offset never travels more than about +/-15mV on either channel (hot or cold) and bias stays put at about 1A85 on both channels once warmed up and measured over a 14h period.

Thanks for the help sorting this out, Zen! 👍
 
Interesting discovery after swapping SITs and putting this amp back in service: the output level is lower than it was before. This amp drives the left and right speakers for the system in the family room (I use the receiver's amps for center and surrounds). I had to dial the channel levels up significantly higher than before (on the order of +5 or +6dB, I didn't make specific notes) in order to get the output level to match the center speaker level again. Granted, the overall bias level is a little lower now, but that shouldn't impact overall input/output sensitivity. With the new SITs in place, I've targeted 0v300 across the PSU chokes that measure 0R16 DCR for bias of about 1A88. Originally, bias was creeping as high as 2A2 (0v354 across 0R16 choke). Bias and output in the amp with replaced SITs seem to be holding steady so far.
 
That's the part that is curious to me. The way that I've understood things (perhaps I'm wrong - wouldn't be the first time), greater bias current should impact maximum power output (higher bias = higher maximum output power). I've never seen bias impact input/output sensitivity before. But, given that the previous pair was already misbehaving to some degree, I guess this shouldn't be a surprise.