Quiescent current question? thanks

i've read a few posts stating that 20 to 25 millivolts across emmiter resistors is a pretty good average. ok, my power amp has 2 mjl4281a/mjl4302 pairs for outputs correct me if i have wrong numbers hehe. i have one dual .22 ohm per channel. my question is, when i measure from mjl4281 emmiter to mjl4302 emmiter for same channel, it's the same as measuring across across the outer legs of the dual .22 ohm resistor for that channel correct? so, measuring across both .22 ohm resistors would be .44 ohm. should pot be set for 40 to 50 millivolts across the .44 ohm instead of 20 to 25 millivolt across one .22 ohm? i have about 23 millivolts across each dual emmiter resistor now .44 ohm both outside legs. this is an Onkyo M-282 2 channel 100 watt per channel into 8 ohm with pretty hefty heat sink. if i develop a lot of transformer hum, am i going to high? even though heat sinks are within normal heat range? thanks, crippledchicken :)
 
richie00boy said:
Yes you are measuring correctly and your assumptions are correct.

23mV across 0.44 ohms gives an Iq of 52mA, about spot on for most transistor output stages.
hi and thanks, could i gain anything by raising the current a little? or mainly circuit dependent. i replaced the pots with 15 turn sealed pot as the others were old and dusty. thanks again
 
amplifierguru said:
Hi Crippledchicken,

You could vary it between 25mA and 75mA and note any perceived audible differences, but it is largely the role of NFB to correct things beyond the nominal Iq setting. The thermal capacity of the heatsink is then not compromised.

Greg
HEY THANKS FOR THE INFO. i was reading the article put a tiger in your amp where they said anything less than 100 ma. wasn't enough and over 150 ma. wouldn't be any improvement. and wanted to get more opinions. thanks again, crippledchicken


:)
 
Hi,
"responsible for a lot of dead and bad-sounding amps"

The Iq will determine the heatsink and transistor temperatures when no signal is output.
Raising Iq will raise the transistor case temperature (Tc) and this will impact on the safe operating area (SOA) of the output transistors.
The designer of a commercial amp will try to optimise the whole package to maximise profit for his employer. This will often result in little margin for running an amp at increased Iq.

However if you are in a cold climate and NEVER experience high temperatures in your listening room then you may have a little leaway for experimentation (but keep a very careful check on Tc and/or heatsink temp). At your level of Iq for BJTs you are unlikely to notice much (any) difference in reduced xover distortion since the designer seems to have chosen a highish Iq already.
 

anatech

Administrator
Paid Member
2004-06-06 8:31 pm
Georgetown, On
I've said it before. Depends on the circuit design. Increasing bias current may not bring any improvement at all. If you want to check, then try by all means. Use a THD meter or better, and your ears. Don't get caught in the trap where you have a few people imagining differences. Like outlets, power cords etc ... That kind of stuff only improves poorly designed equipment at best.

But, if the amp is running too hot, what's better. Getting sound every time you turn it on, or smoke now and again? Don't make your amp unreliable.

-Chris
 
hi fellas, and thanks for all the experienced advice sounds like you know your stuff! i have raised it from 23 mv. across the outer legs of the dual emmiter resistors .44 ohm total to about 30 mv. which i figured to be about 68 to 69 ma. per channel. no excess heat but not much difference in sound though. should i drop back to 23 mv. or should it be safe enough still? also the original drivers have an ft of 8 but, the ones which are suppose to go with the mjl4281a/mjl4302 pairs has an ft of 30 which i have on hand. would this extra bandwidth cause oscillation problems in the driver stages? or, should they be designed to limit the bandwidth maybe? don't know if they would be an improvment soundwise, but i know their suppose to match the output transistors. don't remember the part# on original drivers but do remember the ft of 8. if i do change drivers, i guess i'll have to rebias again correct? thanks again, crippledchicken :)
 

anatech

Administrator
Paid Member
2004-06-06 8:31 pm
Georgetown, On
Hi crippledchicken,

Your Fluke 73 is useless for this - sorry. Good meter otherwise.

What you want to do is measure the DC current gain and match the devices that are the closest. An old Heathkit IM-18 is really good for this. There are many other meters with an hFE scale. You can use a curve tracer, but that is overkill. You just want a simple answer.

Try to rig something up that will put out a constant current into the base of the test device and measure the current. Set up so the measured current is close to your target bias current. You may be injecting 60 to 120 times less than your target current. Use at least 10 VDC C-E voltage. A heatsink will keep the reading more consistant.

Once installed in your circuit, the voltage drop across the emitter resistors should all be close. Just a way to check your own work.

-Chris
 
anatech said:
Hi crippledchicken,

Your Fluke 73 is useless for this - sorry. Good meter otherwise.

What you want to do is measure the DC current gain and match the devices that are the closest. An old Heathkit IM-18 is really good for this. There are many other meters with an hFE scale. You can use a curve tracer, but that is overkill. You just want a simple answer.

Try to rig something up that will put out a constant current into the base of the test device and measure the current. Set up so the measured current is close to your target bias current. You may be injecting 60 to 120 times less than your target current. Use at least 10 VDC C-E voltage. A heatsink will keep the reading more consistant.

Once installed in your circuit, the voltage drop across the emitter resistors should all be close. Just a way to check your own work.

-Chris
WOW! great info really appreaciate it thanks


:D
 
anatech said:
Hi crippledchicken,
You have some work ahead of you. Just for chuckles, build with matched and listen. Then put an unmatched set in and listen. Let us know what you find.
-Chris
hi anatech, i didn't bother matching parts but, i installed the drivers that match the new outputs and did, get bettet mids,highs and dnamic range. i reset quiescent current back to 52 ma. per channel and replaced the op amps with ad8610s with decoupling and film and foil bypass. has only 4 to 5 mv. dc offset so i think i'll call it good. FOR NOW HEHE! CHEERS! CRIPPLEDCHICKEN
 
AndrewT said:
Hi Cripps,
Please finally confirm your drivers, outputs, predrivers (if any) and VAS and confirm the Iq in each.
HI AndrewT the drivers i put with the mjl4281a/mjl4302 are the mje15035/mje15034 pairs. i gues the best thing for me to do would be try to get some type meter to match hfe with on the transistors if their not too expensive as i live on a fixed income. sorry for taking so long to get back to you. reguards, crippledchicken


:)
 

Greg Erskine

Member
Paid Member
2002-01-05 11:56 pm
Sydney
Hi crippledchicken,

I'm not an expert like others that have already posted but If I read your posts correctly you have only 4 output devices for 2 channels. So, are you intending to match the mjl4281a/mjl4302 pairs or the 2 mjl4281a and 2 mjl4302 to each other. Either way, I think it is more important to match your input pair rather than your output pair. Your input transistor are ususally identical type/brand and can be matched.

Reading your initial post, I think you were on the right track. Assuming EF topology which is significant, according to Douglas Self the optimal Vq for minimal crossover distortion across 2 x 0R22 (Re) is 46.2mV. He states he found optimal Vq to be fairly constant (42.6mV to 54.8mV) when changing Re from 0R1 to 0R47. Iq on the other hand changes dramatically because Vq is reasonable constant but Re changes 500%. So my understanding is we should be concentrating on Vq not Iq. Randy Sloan also recommends setting Vq in a similar manner with similar values.

I have tried this on a cheap commercial amp (EF) I have, changing Vq from 27mV to 55mV and it didn't blow up. On my test system I couldn't really notice a difference and I "imagined" the heatsinks got slightly warmer. Not very scientific but at least I know Vq is the same on both channels now.

On my kit amp (EF) I set Vq as recommended to 55mV.

Hope I haven't lead you astray,