[rellik]

how do i go about adjusting the dc offset for the dynahi?

should i set it to zero offset upon starting up or after warm up?

thanks
Hersh Singh

 

[Garbz]

I let mine warm up since it'll be listened to warm not cold.

The startup offset isn't that bad. The initial turnon and defnitly turnoff thumps are though.

 

[steinchen]

I second adjustment after warmup.

Turnon and turnoff thump ? My Dynahi and Dynalo make zero thump, in contrast to opa637 / ad843 chipamps.

 

[amb]

If you have both 10KΩ trimpots installed on each amp board, then you use them to adjust the bias and offset.

First, make sure that the DC servo opamps are unplugged, then power-up the amp and let it warm up. Set the volume control to minimum setting (if equipped). If not, then short the left and right inputs to ground with a 1KΩ resistor.

Measure the voltage across any of the 20Ω output resistors. For the proper amount of bias, you should get 0.75V (which corresponds to 75mA of current on each output transistor branch). You should turn each of the trimpots to increase or decrease the voltage. If the voltage is too high and you can't reach 0.75V even at the minimum trimpot position, then you should replace the 500Ω resistors that are in parallel with each of the trimpots to 620Ω or 680Ω and re-test. Once you get about 0.75V on all the 20Ω output resistors, check the output DC offset. Turn one of the two trimpots to change the offset toward 0mV, a little at a time, and simultaneously turn the other trimpot in the opposite direction. Re-check the voltage across the 20Ω resistors, make adjustments as necessary. Repeat this until you get less than 1mV of DC offset, and still have the correct voltage across all those 20Ω resistors.

Do the other channel the same way.

Wait 15 minutes for all voltages to settle, re-check and readjust slightly if necessary. Once you're done, power off and let all capacitors drain down. Install the DC servo opamps and power up again. Check the DC offset to make sure all is well. Remove the 1KΩ shorting resistors at the input if you had put them on.

Hope this helps.

 

[Garbz]

They don't thump in the true sense of the word. But the DC ramps up when you power it off. It's simply the supply caps discharging. Mind you my headphones in my original and pooly setup dynalo suffered 700mV for a few seconds at every poweroff and still work just fine so i don't see it being too much of a problem.

 

[rellik]

thanks for all of your help

would it be possible to socket another component with less leads since i forgot to socket the opamp in the servo?

 

[amb]

Quote:

Originally Posted by rellik would it be possible to socket another component with less leads since i forgot to socket the opamp in the servo?

Yes, just remove the 5KΩ resistor at the output of the DC servo opamp, and you have effectively disconnected the servo.

 

[grasshpr]

From the original Dynahi construction page, I was under the impression that the DC offset should be least at the turn-on stage, mainly because the slow thermal drift can be accounted for by the servo.

I just finished the electronics of my dynahi and have noticed that if I adjust the trimpots to account for DC offsets near thermal steady state, then the initial DC offsets (when amp is cold) are quite high, 5-10 millivolts. However, if I adjust the initial DC offset when the amp is cool, then at thermal steady state, I usually reach a maximum DC offset of about 3 millivolts.

I guess my rationale is that I don't want to wait 10 minutes for the dynahi to be thermally steady before I actually use the amp.

 

[rellik]

i understand completly now

thanks everyone

 

[amb]

grasshpr, the DC servo, when installed, will correct for DC offset whether the amp is cold or warm. Adjusting the amp for minimum offset when warm (which is the normal operating condition) makes it such that the servo needs to do the least amount of correction in this state. If the transistors are reasonably well matched then the DC offset should be very low at all times. My dynahi never exceeds 1mV of offset, cold or warm.

 

[grasshpr]

Quote:

Originally Posted by amb grasshpr, the DC servo, when installed, will correct for DC offset whether the amp is cold or warm. Adjusting the amp for minimum offset when warm (which is the normal operating condition) makes it such that the servo needs to do the least amount of correction in this state. If the transistors are reasonably well matched then the DC offset should be very low at all times. My dynahi never exceeds 1mV of offset, cold or warm.

I see. Good to know.

Unfortunately I didn't purchase enough trannies to get my system within 1mv offset at startup (only one channel was 1.5mv at startup, the other was around 12mv).

I guess thats a warning to all noobie dynahi builders, BUY LOTS OF TRANSISTORS! These dynahi set of transistors were alot harder to match compared with the dynalo trannies for some reason (besides the SK389 and SJ109's which will cost a fortune to match). I had very little matched pairs from the sa1145/sc2705 and sc3421/sa1358 sets, even from the same gain classes.

 

[Garbz]

You can socket the 1Mohm resistor on the Opamp input too. Heck there's quite a bit you could do to disable it.

Regarding the offset it's not as important as you think. Headphones are considerably less sensitive to DC then people make them out to be. At poweron even with a horribly matched setup (mine

) there's still less then 20mV drift from poweron to running. Headphones can easily handle 5-10 times that before they even start to cook.

 

[amb]

Quote:

Originally Posted by Garbz You can socket the 1Mohm resistor on the Opamp input too.

Actually, no. If you let the input of the servo opamp float while it's powered up (which is what happens when the 1M resistor is removed), the output of the servo opamp is indeterminate (it would probably drift toward one of the supply rails). This will screw up the DC offset big time.

 

[Garbz]

Hmmm works fine at my end. Mind you I'm using OPA134. The output of the servo stayed at 0.

Good to know anyway i'll keep that in mind in the future if I need to diagnose such a problem.