Balanced Phono

[Mike H]

Pages 628 Onward? "Balanced Hybrid RIAA Stage"

[Paul Barker]

It's a little awkward discussing without a link to schematic. Via the link Nick gave you could possibly find the schematic, but I couldn't on my ipad. May be that on a normal computer you can jump to the correct page and not run out of allowable browsing.

My issue is the negative supply for the ccs's. I am wanting to simplify the build by ising lm317's for ccs's. Alan Wright actually suggests those.

Morgan says that the input stage bjt base has to be at 0v so as not to load the cartridge. So he provides a -15v supply and uses it for all hos ccs's. afaik the two further ccs's would not require negative supply, especially with the little headroom required by the 317.

I am having a little difficulty comprehending the need in the bjt. ecc88 cascode. The ecc88 grids in morgan's circuit are grounded. I cant figure what the dc level at the base would be. I am neither sure of the essential need, since the stage is differential and the cartridge connection is balanced with no ground reference at the cartridge. Surely any current across the cartridge would only uccur if the circuit is out of balance? The chosen dual bjt's are meant to ensure balance.

Anyway I am tring to avoid additional complication. I suppose I could pop a negative supply on stage one only just to be sure. No need at all for it in stages two and three as the grids of those stages are hround regerenced at the centre grounded ggrid resistors, and cathode level is forced by the ccs, connection to these two stages and connections out are both protected by coupling caps. As the 317 method doesn't require a negative supply to function, I can't see the problem simplifying this element of the build. Although reflecting, maybe the bias voltage desired at the second valve stage is less than the headroom need by the ccs. Perhaps I best provide a small negative supply. probably -6v would be fine.

[Paul Barker]

I think I just cancelled out everything I had to say in the final sentmese above! The workings of the dyslexic trying to pen a conversation.

[Paul Barker]

I've just avoided a clanger. The mje340 on the final output stage has to deal with 100v. That would have killed my 317. I need to put a power resistor together with it to consume that voltage. It'll be ok elsewhere without a voltage sharing measure.

The output fets are seeing 0.62 watts, and the ones Morgan uses FQP1n50 are more than man enough. But I can only easily source the FQN1n50C, which is rated at 0.89 watts. I think I'll just drop the current 6mA to 5mA which would mean 1/2 a watt. 5mA should drive the balanced cable to the balanced power amp and the grids of it's low capacitance pentode input valve.

200 ohm 1 watt to take away 80 of the 100 odd volts should cover it. Four required.

[Cressy Snr]

As we discussed yesterday Paul, I'm watching with interest.
Makes sense to have a balanced phono stage with our amps.

[Paul Barker]

ordered the stuff from RS which came to £100. just for resistors and caps that is. forgot about a bunch of stuff as per usual.

got the silicon devices via ebay. have the ecc88's.

Now need in addition to signal stuff ordered above to make power supplies. and in/out chassis sockets and more balanced interconnects.

few of my items from rs were 2 needed 5 package quantity. So I'll pass on the spares to you Steve.

I'm using small 100mA 317's which are more than man enough. I tried to establish if low drop out regulators were suitable for ccs, but never in manufacturerers literature do you see a ccs example, and some when you read the blurb specifically state unsuitability. So I've gone with commonal garden 317. I'll just make a -6v rail to more than make up for dropout.

[Mike H]

It's a little awkward discussing without a link to schematic. Via the link Nick gave you could possibly find the schematic, but I couldn't on my ipad. May be that on a normal computer you can jump to the correct page and not run out of allowable browsing.

Actually no had to keep paging down until I came to it ~ the page counter box was helpful tho...

[Mike H]

I am having a little difficulty comprehending the need in the bjt. ecc88 cascode. The ecc88 grids in morgan's circuit are grounded. I cant figure what the dc level at the base would be. I am neither sure of the essential need, since the stage is differential and the cartridge connection is balanced with no ground reference at the cartridge. Surely any current across the cartridge would only uccur if the circuit is out of balance? The chosen dual bjt's are meant to ensure balance.

OK just looked at it again.

OK so both the ECC88 grids AND the transistor bases are at ground potential. The bases each have 562R resistors to ground. (While the grids are connected direct.)

It works because while the triode grids are grounded the cathodes are positive by several Volts, and is all that the transistors need to work.


I can post a screen grab of the diagram unless Nick thinks there's a copyright issue?

[Paul Barker]

It's just that Morgan mentioned in the text something about the matter that puzzled me. If I am puzzled, I might cock up. I'll dig it out again and get back on the matter.

[pre65]

It's just that Morgan mentioned in the text something about the matter that puzzled me. If I am puzzled, I might cock up. I'll dig it out again and get back on the matter.

You could always send Morgan a PM, he is a member here.

[Paul Barker]

I may have imagined it. So far I find the higher bias current of a bjt compared to a fet, but that is negated as a problem when the load resistors are 0.1% tolerance. So that base is covered. I'll keep reading in case there is yet an issue I've passed over.

[Paul Barker]

Another reference but not relevant was that of the cartridge proves not to be channel balanced, varying the ccs current will fine tune balance.

[Paul Barker]

ok i found it.

triode grid is ground. BJT with Vce of 3.3v base is grounded so emitters are at -.7 and collectors will end up at 2.6v So the ccs needs to allow for -0.7v then there is the dropout and any voltage across the bias resistor and a margin for headroom.

So it was nothing to do with the cartridge.

the small 100ma 317 is comfortable sourcing as low as 5mA. the dropout seems to be 3v the bias resistor for 6ma will consume 1.25v we need another 0.7 so a -6v supply gives us 1v margin of safety. Probably safer to make it -7v then.

[Paul Barker]

The dual matched bjt Morgan suggests and the one Alan used are more or less the same. I happened to have ordered the Wright ones MAT02 , but also order the Morgan ones ssm2210 . only two,needed and about £5 or £6 each.

The only difference is the minimum gain, could go less on Morgan's, but the expected gain and just about every other parameter is pretty exactly the same for both.

Neither have yet arrived though.