BUF03's
Hi George, Depending on age, BUF03's maybe labeled as being manufactured by PMI or Analog Devices. The EJ's are common and as George mentions the AJ's are better as far as DC specifications are concerned. As to fakes... I have been lucky that the several I got from Asia were the real deal. The heatsinks are a bit tough to come by and are required.... These suckers get hot! Obsolete however really worth looking for.
Hi George, Depending on age, BUF03's maybe labeled as being manufactured by PMI or Analog Devices. The EJ's are common and as George mentions the AJ's are better as far as DC specifications are concerned. As to fakes... I have been lucky that the several I got from Asia were the real deal. The heatsinks are a bit tough to come by and are required.... These suckers get hot! Obsolete however really worth looking for.
Yes unfortunately there is, a massive one.
The noise specs on AD844 are for it running closed loop as per
traditional opamp with feedback. As such the internal noise of the current
mirrors is largely negated because they are inside the feedback loop.
Once you 'open the loop' the noise at OP becomes a RMS sum of all
these internal noise generators. Running as a grounded base current
conveyor as per your (Georges) application is not too bad because the
OP load resistor is fairly low value. It still may degrade the potential DR
of 1704.
However running 844 open loop with high gain and low impedance feeding it
may subjectively sound great but it is for sure a distortion and noise
generator.
In my past work, grounded base I-V's optimized for distortion and noise
sound better.
Z
Hi George, Depending on age, BUF03's maybe labeled as being manufactured by PMI or Analog Devices. The EJ's are common and as George mentions the AJ's are better as far as DC specifications are concerned. As to fakes... I have been lucky that the several I got from Asia were the real deal. The heatsinks are a bit tough to come by and are required.... These suckers get hot! Obsolete however really worth looking for.
I've used both PMI and the newer AD both mil spec AJ, PMI was a division or bought by AD. They both souned as good as each other, the AD has thinner legs.
The PMI got to 50c with the heat sink I showed in an eariler post, in the same unit with the same heatsink the AD got to 48c on the same day. I think AD probably droped the class A bias just a touch when they started making them.
Cheers George
Zenelectro, I just got this email from Audiogon that you posted here with a sim.
***************
I've done some simulations of the AD844 running open loop.
Hopefully this will help some here understand more clearly how this works.
Here's the basic circuit as configured for a 1704. The DAC OP impedance is 1k,
the load is 2k so we get +-2V swing at the OP.
I've set the IP pair at approx 1.6mA quiescent which I think is probably pretty
close.
THD is about 0.008%. In real life this would be worse but this is a reasonable
indication.
***************
There may also be other replies, but you will not receive any more notifications until you visit the forum again.
All the best,
diyAudio
Is there a problem with this site as I cannot see this post or sim pics.
Cheers George
***************
I've done some simulations of the AD844 running open loop.
Hopefully this will help some here understand more clearly how this works.
Here's the basic circuit as configured for a 1704. The DAC OP impedance is 1k,
the load is 2k so we get +-2V swing at the OP.
I've set the IP pair at approx 1.6mA quiescent which I think is probably pretty
close.
THD is about 0.008%. In real life this would be worse but this is a reasonable
indication.
***************
There may also be other replies, but you will not receive any more notifications until you visit the forum again.
All the best,
diyAudio
Is there a problem with this site as I cannot see this post or sim pics.
Cheers George
You guys are fast!
I pulled the post after doing some component substitutions and realizing that
the major contribution to distortion was the current mirror transistors.
The sim was done at 20kHz so HF distortion due to voltage modulated capacitance is a major player.
I know most you guys get confused by all this stuff so I pulled the post till
I get time to optimize the spice sim to show what is happening a simple
fashion as opposed to exposing the many subtle factors that affect HF
distortion. I understand all this stuff implicitly but it seems not many others
do and as such seem to lose interest when things get beyond a certain
level.
Once you fully analyze these I-V's they are quite complex.
Needless to say I did simulate an 844 running a very low value resistor
at the IP with a lot of gain and guess what - truckloads of distortion as
predicted.
A word to Audiogon - how about PMing ME WRT my pulled post
instead of sneaking around other people
So yes I can continue and show a few different iterations but it will take time
and I'm not going to do it if there is no interest.
cheers
You guys are fast!
It came to me 10hrs before I posted, not fast at all.
Cheers George
Yes, take the output to your RCA's from pin 6, it is very good, but if you want a better buffer than the internal one of the AD844, I would take the signal not from pin 6, but from pin 5 to a separate BUF03 output buffer.
Cheers George
Last edited:
Totally with you there, the best cap is no cap.
I know with the PCM1704 it can be eliminated (dc coupled) as it only gives a couple of mV dc offset which I can null out in the BUF03 buffer with it's dc nulling pot pins 1 and 5. Maybe the same can be done with the TDA1541, as I have not used this dac.
Cheers George
Using TDA1541A with AD844
Yes, You can eliminate the capacitor. You adjust the 2SK170 so that when measuring at the input you get as low a value as possible. Should be a mV or less. The output of the 844 will then be around -1.2 mV (typical). This is an easy direct couple. You can get better results as George says by using a BUF03 across the pin 5 TZ resistor. With that setup I get 0.1 mV offset direct coupled all the way. I use a nulling circuit at the BUF03 also. I also have a discrete version of Pedja's circuit up and running. That did require a cap although I am told there are ways around that too.
Yes, You can eliminate the capacitor. You adjust the 2SK170 so that when measuring at the input you get as low a value as possible. Should be a mV or less. The output of the 844 will then be around -1.2 mV (typical). This is an easy direct couple. You can get better results as George says by using a BUF03 across the pin 5 TZ resistor. With that setup I get 0.1 mV offset direct coupled all the way. I use a nulling circuit at the BUF03 also. I also have a discrete version of Pedja's circuit up and running. That did require a cap although I am told there are ways around that too.
Just out of interest, (if you measured it) what is the dc offset before the coupling cap on Pedja's discrete circuit?
Cheers George
Pedja Discrete offset
Hi George, Due to the 2SK170 current source you'd think it would be possible to null it. I ran into a situation where 10 to 12 mV was the best I could do. The trouble is the circuit is a bit thermally sensitive. So if you blow across the transistors in the current mirror. The offset goes all over the place. The cap was an ends to a means. I think if you matched devices and made a sink to join the transistors so the temp was averaged.... Maybe you could direct couple.
Hi George, Due to the 2SK170 current source you'd think it would be possible to null it. I ran into a situation where 10 to 12 mV was the best I could do. The trouble is the circuit is a bit thermally sensitive. So if you blow across the transistors in the current mirror. The offset goes all over the place. The cap was an ends to a means. I think if you matched devices and made a sink to join the transistors so the temp was averaged.... Maybe you could direct couple.
Caps
Hi George, You shouldn't need a current source with the PCM1704. It really is an issue with the TDA1541A. I certainly agree with you as caps go. Even the Mundorf I tried wasn't that great. I am now using old Relcap Teflon and tin foil. I have had that pair in my parts stash since 1996. First time I could use them. They were expensive in 1996 US Dollars.... Anyway I like the sound. I have enough parts on hand to try matching some transistors for a discrete build. I also have a pair of PCM1704's and a pair of the DF1704's. The plan is to give that a whirl by fall. I am thinking WM8804, DF1704, PCM1704's and a version of the discrete diamond Pedja circuit. I am going modular so I can try a bunch of dacs I have on hand. Time is always an issue.... Dave
Hi George, You shouldn't need a current source with the PCM1704. It really is an issue with the TDA1541A. I certainly agree with you as caps go. Even the Mundorf I tried wasn't that great. I am now using old Relcap Teflon and tin foil. I have had that pair in my parts stash since 1996. First time I could use them. They were expensive in 1996 US Dollars.... Anyway I like the sound. I have enough parts on hand to try matching some transistors for a discrete build. I also have a pair of PCM1704's and a pair of the DF1704's. The plan is to give that a whirl by fall. I am thinking WM8804, DF1704, PCM1704's and a version of the discrete diamond Pedja circuit. I am going modular so I can try a bunch of dacs I have on hand. Time is always an issue....
DaveYes, You can eliminate the capacitor. You adjust the 2SK170 so that when measuring at the input you get as low a value as possible. Should be a mV or less. The output of the 844 will then be around -1.2 mV (typical). This is an easy direct couple. You can get better results as George says by using a BUF03 across the pin 5 TZ resistor. With that setup I get 0.1 mV offset direct coupled all the way. I use a nulling circuit at the BUF03 also. I also have a discrete version of Pedja's circuit up and running. That did require a cap although I am told there are ways around that too.
thanks for the tip, but what is the value of the adjustable VR1 from the schematic? I would be needing also a simple diagram or description in a few words about the dc offset nulling on the BUF03's side in case I decide to try that one too (I saw they are pretty expensive on ebay)
It's all here post 756
First page of the PDF
http://www.diyaudio.com/forums/atta...1397769863-using-ad844-i-v-buf03-original.pdf
5kohm trimer between pin 1 and 5 wiper goes to 7.
Cheers George
AD844 with TDA1541A current source adjustment
Hi, I apologize for the late response. The value of VR1 is 1541A dependent. So I will tell you how to adjust it. First you will need a multimeter with a DC Voltmeter setting with a mV scale (milliVolts). Attach the positive lead to pin 2 on the AD844. The negative lead goes to a ground near the AD844. It maybe necessary to go to a higher scale for the initial adjustment. If so. Use a 2 Volt or 5 Volt scale. Adjust VR1 for 0 volts. Once your there switch to the mV scale and readjust for 0 mV. You may see some settling. It helps to have the dac warmed up. Try and get a value near 0. It isn't necessary to be dead on. The purpose of the current source is to counter the 1541's normal offset. Once you happy with the zero. (No music playing) It is time to check your AD844's offset. Probe pin 6 on the AD844 with the positive lead and the negative lead on ground. Most measure -1.2 mV or so. Repeat the procedure for the other channel. For the BUF03 across the TZ pin 5 resistor. First you do VR1. as above. Once the zero is good on the dac. Measure at the BUF03 output pin. You will have the positive lead on the output pin and the negative lead on the ground. Use the nulling pot. to adjust for 0 mV. I usually can get to about .1 mV. Make sure you heatsink the BUF03. It is heavy class A and will burn up without a sink on it. Reference post #787 for the AD844 circuit schematic. Once your happy with all that consider stacking AD844's. Two are nice. I have a 3 stack. If you try stacking throw me an email and I will tell you which pins to hook up and which ones to float. I hope that helps. Dave
Hi, I apologize for the late response. The value of VR1 is 1541A dependent. So I will tell you how to adjust it. First you will need a multimeter with a DC Voltmeter setting with a mV scale (milliVolts). Attach the positive lead to pin 2 on the AD844. The negative lead goes to a ground near the AD844. It maybe necessary to go to a higher scale for the initial adjustment. If so. Use a 2 Volt or 5 Volt scale. Adjust VR1 for 0 volts. Once your there switch to the mV scale and readjust for 0 mV. You may see some settling. It helps to have the dac warmed up. Try and get a value near 0. It isn't necessary to be dead on. The purpose of the current source is to counter the 1541's normal offset. Once you happy with the zero. (No music playing) It is time to check your AD844's offset. Probe pin 6 on the AD844 with the positive lead and the negative lead on ground. Most measure -1.2 mV or so. Repeat the procedure for the other channel. For the BUF03 across the TZ pin 5 resistor. First you do VR1. as above. Once the zero is good on the dac. Measure at the BUF03 output pin. You will have the positive lead on the output pin and the negative lead on the ground. Use the nulling pot. to adjust for 0 mV. I usually can get to about .1 mV. Make sure you heatsink the BUF03. It is heavy class A and will burn up without a sink on it. Reference post #787 for the AD844 circuit schematic. Once your happy with all that consider stacking AD844's. Two are nice. I have a 3 stack. If you try stacking throw me an email and I will tell you which pins to hook up and which ones to float. I hope that helps.
Dave
Once your happy with all that consider stacking AD844's. Two are nice. I have a 3 stack. If you try stacking throw me an email and I will tell you which pins to hook up and which ones to float. I hope that helps.
georgehifi said: