Hi,
After hours of debugging, I finally found the problem. It was the stacked AD844 causing instability. Tried more bypass caps (47uf+0.1uf each rail) but not helping at all. I am curious why it just worked flawlessly yesterday. Put one AD844 in the socket and the output DC reduced to 0.1mv right away! (a bit more with input) Maybe the AD844 is not stackable?
Hi Torchwood421
a little off topic, how would you rate the Edcor transformers compared to the Lundahl for sound quality?
AD844's
Could be you have latent damage from that previous experiment with all pins soldered.
Last edited:
Input Transformers...
For most people with good systems the Edcor's are very inexpensive and they behave well. Lundahls are crazy expensive. I have used the less expensive and equivalent Cinemag's with just as good results. Jensen's are also very good although I have not personally tried them. If you have a very high end system it pays to use at least Cinemag level parts. Just my opinion. 😉
Hi, the instability occurred around 70% of power cycle. Tried 3 stacked instead of 4, and the chance of instability lowered. I read somewhere in this forum that each PCM63PK from parallel config needs separate bypass caps in order to work correctly. Maybe this also applies on AD844. I will do some experiments and see what happens.
Last edited:
This looks like a interesting buffer to make for the 844 using dual maybe opamps for ease.
I would though use fet input opamps so the 844's TZ output has higher impedance to look at and change the input so the TZ sees 200k, load.
http://www.diyaudio.com/forums/analog-line-level/155143-class-aa-opamp-schematic.html#post4654086
Cheers George
I would though use fet input opamps so the 844's TZ output has higher impedance to look at and change the input so the TZ sees 200k, load.
http://www.diyaudio.com/forums/analog-line-level/155143-class-aa-opamp-schematic.html#post4654086
Cheers George
Attachments
Thanks, Lundahl makes me confused as they have some many model with similar purpose, half a dozen models of 600:600 line level transformers. Some of them are cheaper than Cinemag. the armophous core type is very expensive though. The edcor rolls off by around 0.75db at 20khz so I think it may be more suitable for Vout delta sigma dac. NOS dac already has slightly rolled off high so it should be better to use with wider bandwith transformers like Lundahl?
George
I recognize the ebay seller name in the schematic. If this is still available then I think we can get it ready made for very reasonable price from ebay.
Cheers
The pcb for this one is available on ebay for 4 bucks including shipping 🙂. It seems that many people have bought it.
Trafo's...
This is of course true. They have a number of models. I can only speak about the LL1690's. As I own a pair. They are the more expensive ones with amorphous cores. The Edcors did work fine after my I/V and buffer although I stuck with the XSM600/600 in 2.5 Watt form. No chance of saturation with such large EI cores. Didn't do much testing however 0.75 db down at 20 Khz is pretty good and where you want the roll off to start. My cousin is using a pair in his CS4397 DAC. A Delta Sigma dac. I can say they sounded excellent in the CS4397 dac. That is the WSM series. I bought a pair to use with my CS4397 build, although I have not tried my WSM transformers. NOS.... In NOS use, I don't think you gain much even with compensation there isn't much beyond 15 Khz to preserve. Most people don't realize that there are other options to the SAA7220 digital filter for use with the TDA1541 series dac's.
I bought one, for $4 WTF. I'll get around one day to hear it. Big class A doubles the output current, got a lot going for it. Wonder how it sims, I asked Art Vandelay he said he would sim it
Cheers George
Hi
What do you recommend for the power supply for ad844 and the buffer? does it help much with the performance if the PS is tototally independent from the DAC module?
Is It worth to invest in ultra low noise regs?
What do you recommend for the power supply for ad844 and the buffer? does it help much with the performance if the PS is tototally independent from the DAC module?
Is It worth to invest in ultra low noise regs?
Power Supplies
I am using LT1083 based regulators from Jim's Audio on Ebay. Good quality stuff. I configure these as both positive and negative regulators. 😉 I also have dedicated analog power supplies separate from the DAC digital supplies.
I have two LT1963 but to use them I will need one more power transformer. LT1963 chip has fairly low noise but ripple rejection is only average (not better than lm317, I think). to improve ripple rejection I am thinking about having a 7815 as pre-regulator. This requires a bit more work.
My tda1541 is supplied by discrete regulator with buried zener ref and mosfet power transistor. when I use this to share with the i/v stange the sound is quite good but I have nothing to measure noise, just a bit concerned about interference between DAC board and i/v stage
My tda1541 is supplied by discrete regulator with buried zener ref and mosfet power transistor. when I use this to share with the i/v stange the sound is quite good but I have nothing to measure noise, just a bit concerned about interference between DAC board and i/v stage
Last edited:
Maybe the LT1963 is different, but I did not get good results from TPS7A4701/3301 ultra low noise LDOs. They usually sound crystal clear, but not smooth, rich or relaxed even with Silmic or some audio output caps. IMO, ultra low noise LDO is best for digital circuit. Just my 2 cents.
canvas, thanks for sharing. I also doubt the effect of those low noise regs in analog section. I will stick with the LM317/337 with a separate transformer for now.
Hi,
Problem solved. There is no damage at all. It appears that a 100 ohm resistor after TZ pin will make everything working smoothly. I haven't lowered the value yet, but it seems 100 ohm is adequate here. Currently, I am stacking 4 chips per channel, and the sound is excellent with deeper stage which never had before. I appreciate Pedja and anyone here who devoted for this great circuit. BTW, I have a source who can supply inexpensive "used" AD844AN (~$1 ea). If you don't mind having used parts, pm me.
Last edited:
So, TZ pins should not be directly connected between stacked AD844 's but over 100 ohm resistors like George suggested before for output buffer pin #6?
No, only one 100 ohm resistor after stacked TZ pins is sufficient. I will try to lower the value, but now I only got few 100 ohm high quality resistors.
Last edited:
This was the way Arty did and sim'ed it.
Yes if using the 3 stack AD844's 3 x pin 6 buffer output, each pin has to have a 100ohm on it separately, then all 3 other ends of the 100ohms resistors go together and that's your output.
Cheers George