The output impedance is very high at pin 5 with my 4.7kohm TZ resistor , so high that a test 100kohm loading resistance drops it's output quite a bit. I say the input impedance of a buffer stage so as not to load it down should be over 1mohm.
I posted all these findings quite a few pages back.
Cheers George
I posted all these findings quite a few pages back.
Cheers George
Triple stack VS Pedja Discrete
Hi all, To date my reference has been the Pedja DDNF circuit in my Mark II TDA1541A dac. Tonight I removed the GIC filter from the Mark I. Installed the Aya II stage. Listened to a single AD844. In a few words open, clear, and more then a little harsh. I finished the 3 stack AD844's SOIC version. Installed it. To be clear I haven't added the BUF03 stage yet. My triple stack has taken on the smooth clean sound of the discrete circuit and then some. The dynamics are impressive and the bass is outstanding. The dac is seeing 100 K Ohms on the Pass BA-3. Works splendidly. I now suspect that the discrete circuit needs the BUF03 too.
Didn't expect that. Looks like more modding is in order. Glad I still have a few spare BUF03 set.... George this is an awesome thread.
Hi all, To date my reference has been the Pedja DDNF circuit in my Mark II TDA1541A dac. Tonight I removed the GIC filter from the Mark I. Installed the Aya II stage. Listened to a single AD844. In a few words open, clear, and more then a little harsh. I finished the 3 stack AD844's SOIC version. Installed it. To be clear I haven't added the BUF03 stage yet. My triple stack has taken on the smooth clean sound of the discrete circuit and then some. The dynamics are impressive and the bass is outstanding. The dac is seeing 100 K Ohms on the Pass BA-3. Works splendidly. I now suspect that the discrete circuit needs the BUF03 too.
Didn't expect that. Looks like more modding is in order. Glad I still have a few spare BUF03 set.... George this is an awesome thread.
Dave great to hear your reaping the benifts of the stacking of the AD844's i/v stage, something special happens with this stacking, and even against Pedja's discrete version I am impressed.
And you know that your loading down the TZ output a bit with the BA-3 100k input impedance load, the BUF-03 buffer will be great fix, (BTW how are you controling the volume?).
Cheers George
And you know that your loading down the TZ output a bit with the BA-3 100k input impedance load, the BUF-03 buffer will be great fix, (BTW how are you controling the volume?).
Cheers George
Hi George, I am still using pin 6 until I get around to building the buffer on TZ pin 5. The stacking sounds great although I use just the bottom AD844's output stage. I am still a bit behind. Do you have a source for the BUF03 boards? The dynamics are superb and the bass is powerful. I am sure the BUF03 will add some refinement. Very happy with it so far.
Yes the BUF03 will be great, even better, and you can adjust out any dc offset, just make sure you get the TO-99 heatsinks for them as they run very hard classA bias. I got some more AJ's (milspec) from here and they worked fine.
BUF03 | eBay
Cheers George
BUF03 | eBay
Cheers George
Absoute max on the data sheet is +-18v, and this would be at normal room temps of around 22c-24c.
Is the top one getting warm? As the one on the bottom will be much warmer.
Cheers George
"And "hi-fish sound" indeed too often has something to do with stability issues, or HF mess, and alike."
I have been wondering about that question. I have experienced many of the sort,but I am a little reluctant to add this one to that account. This modification is so thoroughly tested by Georgehifi and the sound is so stunning that I tend to believe that something is absolutely right here. I am not able to catch anything on my 100 Mhz scope
I have been wondering about that question. I have experienced many of the sort,but I am a little reluctant to add this one to that account. This modification is so thoroughly tested by Georgehifi and the sound is so stunning that I tend to believe that something is absolutely right here. I am not able to catch anything on my 100 Mhz scope
Just a report on the owner I put a triple stack on his cdp with PCM1704K dacs, that has the Wilson Watt Puppy 8's and the Halcro DM68's.
He rang me and so ecstatic was he with what he heard with the differences of 1 vs 2 and now 3 stack 844's, he want's me to go 4 stack, I'll be doing this next week and I'll report his findings.
I said to him this is new territory and it may not do anything better or it may well go backwards, or heaven forbid (for the anti stacking brigade) get even better again.
I know I'll have to monitor the heat of the PCM1704K's but there's no max case temp mentioned in the data sheet.
Cheers George
He rang me and so ecstatic was he with what he heard with the differences of 1 vs 2 and now 3 stack 844's, he want's me to go 4 stack, I'll be doing this next week and I'll report his findings.
I said to him this is new territory and it may not do anything better or it may well go backwards, or heaven forbid (for the anti stacking brigade) get even better again.
I know I'll have to monitor the heat of the PCM1704K's but there's no max case temp mentioned in the data sheet.
Cheers George
Heatsink
Might not hurt to mount a heatsink on top of the PCM1704. I do this with my 1541A S1 Crowns using special heatsink tape made for video cards. I'm trying to make these last for as long as possible. Please report back on the 4 stack. I just got the BUF03 circuit boards in the post. So will give that a go next week. Dave
Might not hurt to mount a heatsink on top of the PCM1704. I do this with my 1541A S1 Crowns using special heatsink tape made for video cards. I'm trying to make these last for as long as possible. Please report back on the 4 stack.
I just got the BUF03 circuit boards in the post. So will give that a go next week. DaveI am certain that some bright head inside this community will find an explanation. I am also certain that there is an obvious technically cause for this massive improvement in sound quality. I am just an outdated digital man that has been maintaining computer systems when an 80 Mbyte hard-disc took two men to carry ? So I will be not much of a helper here.
Last edited:
The reason for perceived improvement in sound quality when paralleling the AD844 is because the AD844 has low quiescent current , the entire chip consumes only 6.5mA ,and this is very low , so 3 chips in parallel will be better , as they will share the a.c. Current from the dac chip. another aspect that can be improve by paralleling the 844 is the input impedance , if you use 3 chips the input impedance will be 3 times lower. So in theory the more chips you parallel the better, but in practice it can bring problems , like oscillations.
One must understand that this chip was not designs to this application, but none the less it works fine.
I must say that I am curious to test the 844 topology in a discrete circuit. I have already made some simulations in a circuit with the exact topology of the 844 but with a much higher quiescent current and a much lower input impedance (2ohm) than the chip, the tests show a very good linearity, and even that I already have a lot of I/V converters, I will probably build this one also.
Last edited:
The lowering of the input impedance is easy to understand, but as you point out Sergio and as well georgehifi has pointed out several times the quiescent current must be a factor to be taken into account. It sounds to me that the chip is not able to deliver the current consumption needed on demand from the DAC. What confuses me is the past DAC designs implementing one AD844 without detecting this serious flaw?
Just to give the readers of this thread an impression of what this is like. I have listened to my dac with the impression that this is as good as it gets for, I am not shure, over ten years. And here comes georgehifi and tells me you have been fooled, and I have. I am now listening to my CD collection as I have not listened to them before. Fantastic, I am in the concert hall. And I know what the concert hall sounds like.
Last edited:
Hi Sergio, s3tup has done a discrete version maybe you could sim it, and Pedja Rodic has done it as well and posted up somewhere, maybe someone has it and can direct you to it or post it up as well. But I think we are barking up the wrong tree with the current thing in my post above this.
Here is s3tup's one he did.
AD844/PA630A knockoff: Discrette.
Cheers George