I installed the dc offset trimming circuit with 20k trim pod. It works but temperature drift is large. The diffrent between cold start up and warmed up is more than 40mV. I am still using the internal buffer of ad844 chip. If I trim to near zero at Tz pin with open chassis then when I close the chassis and play for a while it drift more than 20mV.
Could it be that the ad844s have to work hard driving the preamp therefore the drift? I hope this drifting would reduce when I put in an ext buffer.
Could it be that the ad844s have to work hard driving the preamp therefore the drift? I hope this drifting would reduce when I put in an ext buffer.
AD1862
I'd doubt it. Mine was totally stable. The wild card is the AD1862. It should probably be measured cold to warm. Then again I have breathing room between all of my AD844's. Try measuring the dac itself before blaming the AD844 stack. If nothing else you will have a little more to go on. Are you sure it isn't oscillating? I haven't had a chance to work on my DDNF for the PCM1704 this week. After cleaning flux of the board I did better the offset slightly. Even better with rails set to + and - 15 VDC as opposed to + and - 19 VDC. Compliance is 0-1 mA with the 1704 so I don't think it is an issue.
The 844's with or without buffer, don't drift once zero'ed. It's something else.
Cheers George
Cheers George
The offset of AD844 is trivial. 40mv is way too much.
The trimming function is not recommended by ADI anymore. For every mv you trimmed, it could add several uv per °C. This function is for legacy circuits.
The trimming function is not recommended by ADI anymore. For every mv you trimmed, it could add several uv per °C. This function is for legacy circuits.
Last edited:
Where have you read this?
I trimmed 20mV one channel and 15 for the other with no heat problem on a 3 stack.
Cheers George
Hi George,
If I remember correctly, one was from datasheet of ADA4627. Trimming is useful, but it looks like new generations do not support trim function anymore, e.g. OPA827. BTW, my 5 stacked shows 5mv offset after external buffer without trimming. However, the supply is +-5V.
Poting
Last edited:
In my case the AD844 stack was also very low by it's self, it was the outputs of the PCM1704's that had the 15 and 20 mV of offset.
Cheers George
Cheers George
my power supply uses lm7812 and lm7912. the +/- rails are slightly different from 12vdc. I have a filter cap 330pf at the output of the ad844. Maybe there is a problem with the capacitive load. Pedja schematic for early version of tda1541a DAC spec up to 1nf at the output of ad844. But the datasheet says max cap load is about 100pF. I will check again tonight.
I am referring to the capacitors CO1 and CO2 in this schematic.
http://lampizator.eu/LAMPIZATOR/TDA1541 corner/1541 shema.pdf
This looks like 300pf capacitance in series with 75 ohm resistor to ground. Could this cause the op amp to oscillate?
Offset nulling on the dac output
The dac offset 2 mA current source was meant for the TDA1541A which has a very large offset. You really should not need that with a AD1862. I have seen drift with the TDA1541. So don't count on it as having long term stability. 2SK170 has more drift then a BF245A in my experience.
Mine is actually tda1541a. I have just changed the mkp decoupling caps with the Pana PPS caps as you suggested. Sounds really nice. I put 4x 0.1uF on the MSB pin.
TDA1541A Offset
Hi, Actually they both have high offsets be it the early TDA1541 or TDA1541A. The "A" chip is just more refined. I enjoy the effortless dynamics of the not A chip. PPS is hard to beat. Smooth detailed and a joy to listen to. Even better once you get some time on it. 🙂
There isn't much choice for good sounding caps on a tight space. I got a bunch of c0g/np0 0.1uf~0.22uf ceramics, and they do sound very nice. Very neutral with good bass. From your recommendation, I think I will try the Pana PPS.
c0g/np0 caps are very good but the value you mentioned seem to be quite large for this type. I haven't seen such high value but I haven't look very hard for them yet.
Yes, they are not common, but still available on most online stores. I have a reel of TDK 0.22uf. Perhaps I will sell some on Swat Meet section.
Over here http://www.diyaudio.com/forums/lounge/146693-john-curls-blowtorch-preamplifier-part-ii-8076.html someone had an RIAA preamp design where a manufacturer substituted in X7R ceramics caps in place of the specified COG ceramics - there was a 48dB increase in THD!
Important to note the location in the circuit that the X7R was placed in - its part of a filter network, not a coupling cap.
@canvas - what case size are your 0.22uFs? My guess is 1210? I have a reel of 1206 100nFs, for higher values I put them in parallel as the price rises fairly steeply above 100nF.
@canvas - what case size are your 0.22uFs? My guess is 1210? I have a reel of 1206 100nFs, for higher values I put them in parallel as the price rises fairly steeply above 100nF.
I believe they are 1812 in size.
For X7R, I used them (1uf in my case) on my OPA861 IV, but the sound is harsh. I probably won't use X7R on analog circuit again. IMHO, the C0G/NP0 is a totally different world, although they are both called ceramics.
poting