Help me decide which ebay dac please?

[Andrew]

Ok, I will, thanks Nick. Listening further with the new board it is indeed still distorting with full output from the Touch.

Is this the 93 or 96 based board Ali?

Could the valves have drifted and caused a DC offset on the output TX?

Anyway, here's some useful news, I got mine going

I didn't add the supplied bridge rectifier chips as I intend to build another set of off-board regulators and shove DC directly in. I am current running off bench supplies, one set of +15v/-15v DC and the other +12v DC, it seems OK heat-wise; I'll get the temp probe out later on.

Just for the record the +15v side draws 55mA the -15v draw 15mA and the 12v supply doesn't even register on the supply's meter which is not sensitive enough @ 100mA per div. I can put a meter in line, if anyone is interested and will probably do that so I can size the supply accordingly. Since I'm doing this I may take the +12v supply via a reg off the +15v as a reg chip is cheaper (and smaller) than a separate TX.

I will measure the DAC at the weekend, but it sounds very good, better, or at least as good as the SB3, even with the stock op amp in and doesn't seem to be distorting.

One little foible I noticed is it seems picky which supply comes up first, one way around sweet music, the other it just pops and whistles. I think the 9v reg has to come up first but I'd need to check this to be sure.

I changes the reg chips for better ones, upgraded the caps around the op amp and on on the receiver and used PRP resistors for the low pass filter and IV stage, otherwise its stock. I used the LPF for the 96 as per the spec sheet.

That's about it.

Andrew

[Ali Tait]

Andrew, after taking the dac to Nick's to measure on his scope, it would appear that the Sowter trannies cannot tolerate the dc offset they say they can, and this is causing the distortion. Nick biased the valves to match them up more accurately using his sig gen and it seems to be working ok at the moment. I believe Nick is chatting with Bud Purvine about having some more suitable trannies made, so hopefully that will sort the problem.

[Ali Tait]

BTW, I bought some 60's gold grid 6n6p's from ebay, and they do appear to sound better than the ones I have.

[Andrew]

Andrew, after taking the dac to Nick's to measure on his scope, it would appear that the Sowter trannies cannot tolerate the dc offset they say they can, and this is causing the distortion. Nick biased the valves to match them up more accurately using his sig gen and it seems to be working ok at the moment. I believe Nick is chatting with Bud Purvine about having some more suitable trannies made, so hopefully that will sort the problem.

Oh, just as I bought a pair of these to copy Nick's circuit

I wonder if I need to fit a meter like Max's to check the balance?

Nick, how far out were they? Would a CCS help?

Andrew

[Nick]

Its a combination of things. The 20hz or so saturation point is about 2.2v, if the DC offset is added to that the nickel saturates quickly.

The simple way around it I think is to use a 4dB or so pad between the input and the 6n6p grids. I thinks its only a issue about the level at which the core saturates.

[Andrew]

Its a combination of things. The 20hz or so saturation point is about 2.2v, if the DC offset is added to that the nickel saturates quickly.

The simple way around it I think is to use a 4dB or so pad between the input and the 6n6p grids. I thinks its only a issue about the level at which the core saturates.

What would you say would be the best way to achieve this 4db drop?

Andrew

[Nick]

Inline resistor between the DAC and the grid. Resistor to ground from grid. Repeat for each grid.

[Andrew]

Dat's what I thought you would say

I have grid stoppers already, but I wonder how much drop I will get across these as a volatge divider with the grid's internal resistance, probably not that much. Hence the need for a shunt too, I guess.

Andrew

[Andrew]

By the way, the +12v has to be up before, or perhaps at the same time, as the +9v or mine seems to sulk. Receiver coming up before the DAC, or the other way around, perhaps?

Andrew

[Andrew]

I thought this might be a nice output stage, assuming you didn't want a balanced output.

http://www.tubecad.com/2011/03/blog0203.htm

Andrew

[Mike H]

Ey up an op-amp dun wiv valves

[Laurence]

Andrew, after taking the dac to Nick's to measure on his scope, it would appear that the Sowter trannies cannot tolerate the dc offset they say they can, and this is causing the distortion. Nick biased the valves to match them up more accurately using his sig gen and it seems to be working ok at the moment. I believe Nick is chatting with Bud Purvine about having some more suitable trannies made, so hopefully that will sort the problem.

I don't understand the DC offset. The output of the DAC is AC, but it is 2.5v above ground. If it never sees ground there is no DC offset. Why would it have a ground reference if it is driving a transformer?

If you mean that it can't deal with the AC voltage I can see the purpose of Nick's fix of an attenuator.

Nick, what is the ideal primary L of a transformer to load this DAC please? You and Bud must have some idea you are working toward.

sorry I have just realised your problem, the Sowters cannot tolerate 2.5 volts between primary and secondary right? Frightening!

[Nick]

No, the DC offset I refer to is the difference in idle current through the two arms of the primary. The net effect is to saturate the core.

the problem is not a lack of inductance, its a lack of B headroom.

[Laurence]

Ok so contrary to expectations output isn't fully balanced? So are you looking at a much larger core? When I asked what primary L I was trying to establish what load impedance you plan? Nothing to do with B Max.

[Nick]

No, as fully expected the two triodes in the valve are not perfectly matched. Its nothing to do with the DAC output thats perfectly balanced.

The load impedance the valves see is set by a resistor across the output, it works out in the order of 25k A-A.

the problem we were describing is all to do with flux levels, and the saturation point of the core. So yes either a larger core, or one with a less brickwall saturation.