OT personal remarks removed. Please read the forum rules. Personal attacks are not allowed.
Good. 
Let's stick to Electronics
And reminding that all suggestions are offered in good faith, trying to solve a problem.
Fellow Members can't be accused of "not caring", "not collaborating", etc.
No money is involved here, we do it for FUN, on our free time
Back to the OP circuit problem, useful suggestions have been offered; of course the peculiar type of Op Amp specified does add its own complications.
Being a Video specific one, low offset and input current have been sacrificed to extraordinary bandwidth , because, well, in video circuits they are not that critical, since usually AC coupling is acceptable, but most important, such circuits usually are fed from and drive impedances which are incredibly low for our standards (50/75/300 ohms )
Pretending excellent performance when driven from, say, 10K impedances, is not what the designers intended .
Also mentioned before is that such bandwidth will probably complicate physical layout, including grounding, decoupling, and signal routing because of possible instability.
I would have chosen another Op Amp, but hey, to each his own
Let's stick to Electronics
And reminding that all suggestions are offered in good faith, trying to solve a problem.
Fellow Members can't be accused of "not caring", "not collaborating", etc.
No money is involved here, we do it for FUN, on our free time
Back to the OP circuit problem, useful suggestions have been offered; of course the peculiar type of Op Amp specified does add its own complications.
Being a Video specific one, low offset and input current have been sacrificed to extraordinary bandwidth , because, well, in video circuits they are not that critical, since usually AC coupling is acceptable, but most important, such circuits usually are fed from and drive impedances which are incredibly low for our standards (50/75/300 ohms )
Pretending excellent performance when driven from, say, 10K impedances, is not what the designers intended .
Also mentioned before is that such bandwidth will probably complicate physical layout, including grounding, decoupling, and signal routing because of possible instability.
I would have chosen another Op Amp, but hey, to each his own
All good points as I didn't know how video op amps differ from other op amps.
As to bandwidth, can't that be limited with a small cap across the feedback resistor?
I realize I'll have to design a board with ground planes, decoupling caps on the supply pins, etc..
Will using the single version(AD847), which has the offset nulling pins, help with DC offset?
Thanks...
As to bandwidth, can't that be limited with a small cap across the feedback resistor?
I realize I'll have to design a board with ground planes, decoupling caps on the supply pins, etc..
Will using the single version(AD847), which has the offset nulling pins, help with DC offset?
Thanks...
Groundplanes aren't necessary IME, careful decoupling though most certainly is. A 50MHz opamp hardly has 'extraordinary' bandwidth so don't be put off by that. The offset nulling pins are intended to be used for the function of nulling out the internal offset voltage of the opamp and are definitely not recommended for being used to null out offsets from other sources (like unequal input impedances at the two input pins). So the answer to your last question is a 'no'.
OK, so I can forget about the nulling feature. Thank you for explaining what it actually does and doesn't do.
So in reality, can use you such an op amp like this without a capacitor on the output, or is it possible to use say .1% tolerance resistors on both inputs to reduce offset to an acceptable level?
So in reality, can use you such an op amp like this without a capacitor on the output, or is it possible to use say .1% tolerance resistors on both inputs to reduce offset to an acceptable level?
High precision resistors will help but not by much. Let's say you're already using 1% so the variation is 100R. Multiplied by the typical input bias of 3.3uA this is only 0.33mV which is lower than the typical input offset.In the end you'll be limited by the input offset cirremt (50nA typical so it'll contribute 0.5mV with 10k) and the input offset voltage (0.5mV typ). What's your acceptable output offset?
With 10k 1% resistors and the AD827 you can arrive at a typical offset in the region of a mV or so. Not something I'd worry about.
Going down to 5k helps only a little, typically.
If I was worried about 1mV or so offset I'd use a coupling cap. As I previously mentioned, supply caps IME are much more of an issue than coupling caps.
Going down to 5k helps only a little, typically.
If I was worried about 1mV or so offset I'd use a coupling cap. As I previously mentioned, supply caps IME are much more of an issue than coupling caps.
Wow! Just 1mV or so? That's way less than I expected. For some reason I had it in my mind that the offset was going to like be 20-30mV.
I soldered up a AD747(single version of the AD827) circuit this evening(before reading your reply about the nulling circuit not helping with DC offset) using 5K resistors in the feedback loop and from the + input to ground.
I used a 20uF electrolytic from Rg to ground and some small axial lead .1uF ceramics from each power pin to ground.
When I turned it on, much to my surprise, I didn't even measure 1mV at the output with the input shorted.
As you stated, turning the 10K trimmer on the nulling circuit didn't seem to change anything at the output.
I also placed a 2.2uF MKP cap between the 10K volume pot & + input to help eliminate the dreaded "swishing" or "cracking" sounds when rotating the volume pot.
I soldered up a AD747(single version of the AD827) circuit this evening(before reading your reply about the nulling circuit not helping with DC offset) using 5K resistors in the feedback loop and from the + input to ground.
I used a 20uF electrolytic from Rg to ground and some small axial lead .1uF ceramics from each power pin to ground.
When I turned it on, much to my surprise, I didn't even measure 1mV at the output with the input shorted.
As you stated, turning the 10K trimmer on the nulling circuit didn't seem to change anything at the output.
I also placed a 2.2uF MKP cap between the 10K volume pot & + input to help eliminate the dreaded "swishing" or "cracking" sounds when rotating the volume pot.
Last edited:
Sorry, I got "not recommended" mixed up with "won't work" in my head.
After connecting my fiber optic television receiver to the circuit, I found out that the offset circuit does work to reduce offset. After connecting the receiver, I had an offset of around 6mV. With a turn or two of the trimmer, offset came down to like .1mV.
The sound quality seems to be rather flat and one dimensional. Perhaps having a buffer after the AD847 might improve the sound quality. I don't know.
After connecting my fiber optic television receiver to the circuit, I found out that the offset circuit does work to reduce offset. After connecting the receiver, I had an offset of around 6mV. With a turn or two of the trimmer, offset came down to like .1mV.
The sound quality seems to be rather flat and one dimensional. Perhaps having a buffer after the AD847 might improve the sound quality. I don't know.
Yes, one of my sources is the set top box, not sound taken directly from the TV itself.
You're right, it does have a optical output on the back. I have never tried another DAC with the receiver.
I meant that compared to other preamps I've tried, the sound is pretty lacking. Guess you can't expect too much from a single op amp circuit.
You're right, it does have a optical output on the back. I have never tried another DAC with the receiver.
I meant that compared to other preamps I've tried, the sound is pretty lacking. Guess you can't expect too much from a single op amp circuit.
I doubt very much that the poor sound is down to the AD847 (or 827) rather some implementation detail. It already addresses one of the main reasons opamps normally give poor sound - the LTP linearity. Perhaps you've not optimized the grounding or reduced supply noise sufficiently, these issues can often be a reason for flat sound. What power supply are you using?
I find that the sound output from video boxes is quite good.
It's the TV (the bit with the screen) that compromises the audio circuits cf. the video circuits.
My conclusion is consistent with VHS tapes in HiFi record mode and in especially in straight through mode (i.e. as a preamp) and DVD video players and digital set top boxes and digital video recorders (both DVD and HDD types)
Whereas the headphone output from a TV is usually atrocious.
It's the TV (the bit with the screen) that compromises the audio circuits cf. the video circuits.
My conclusion is consistent with VHS tapes in HiFi record mode and in especially in straight through mode (i.e. as a preamp) and DVD video players and digital set top boxes and digital video recorders (both DVD and HDD types)
Whereas the headphone output from a TV is usually atrocious.
I'm using this power supply:
Power Supply Board Kit PCB Based on LM317 LM337 IC | eBay
I know it's probably not a very good supply compared to what I've other members build and use, but I've used it with a lot of other designs(both op amp & discrete) without any issues.
I recently built another single channel preamp design, that another member here posted a while back, using the LT1363 biased into Class A with a CCS. It sounds a little better than the AD847.
The old AD744/AD811 Jung design sounds better, to my ears, than either one of the single op amp designs above.
Ah its a faily 'stock' regulated supply no provisions for HF noise reduction (like rectifier snubbers or passive filtering prior to the regs). But if you also used this supply with the other two circuits and they sounded better I can suggest at least one reason why - supply cleanliness. Both the other preamps you mention will be better in respect of the supply noise - the LT1363 because classA biassing cleans up the supply noise, the Jung composite because the AD811 creates the noise and hopefully the AD744 is separately decoupled.
I'm sure you can improve the AD847 but you'd have to take fairly heroic measures on the supply. You'd also do well to add the CCS, as for the LT1363.
I'm sure you can improve the AD847 but you'd have to take fairly heroic measures on the supply. You'd also do well to add the CCS, as for the LT1363.
I mean the power supply local to the AD847 - i.e. on the board whose schematic you've been showing. Having long leads to an off-board PSU isn't really ideal from a SQ point of view, the lowest impedance needs very tight loops for the current to flow in. So I mean beefing up the local decoupling by a significant margin.
Modding the eBay PSU tp reduce the HF noise its passing can't hurt but its not the lowest hanging fruit.
Modding the eBay PSU tp reduce the HF noise its passing can't hurt but its not the lowest hanging fruit.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.