audio-talk

Posted: **Sun Feb 19, 2023 4:53 pm**

Oh well done Dave. I know I saved it myself but god knows where I put it !

Edit : just threw a page ( again ) ... so annoying. Here's the circuit again -

: LCR-Phono-1.jpg (90.68 KiB) Viewed 9003 times

Posted: **Sun Feb 19, 2023 7:45 pm**

My search foo is good today! Though I see some battles with Spice on the horizon...

Thanks Mark - that image is a bit clearer. The only thing I still cannot make out is the value of the two caps in parallel going to the three pin connector from the 5687 cathode (which I presume is the interface to the RIAA).

Posted: **Sun Feb 19, 2023 8:05 pm**

Yes. Two 33uf films

Posted: **Mon Feb 20, 2023 1:18 pm**

Thanks Nick.

Well, this is frustrating. I've laid out the first stage in Spice so that I can study the circuit and work out what's going on. Unfortunately, I'm getting next to no gain (in fact it's a loss), there's almost nothing on the 5687 cathode. The current going through the first two MOSFETS is on the nA / pA scale and I think that might be the issue. Spice doesn't have IRF740 in its library of N channel MOSFETS so I've been trying to find the nearest match in an effort to get the thing up and running at the very least, even if the sim wasn't 100% accurate. I've tried several but same result. Maybe the MOSFET models are junk.

I've double checked the circuit several times and I don't think I've got anything wrong...

Posted: **Mon Feb 20, 2023 1:29 pm**

What you have missed is the line that says that the depletion mosfets are IXTP08N100's I just used the 740 models as they have the same pinout. The mosfet that's driving the d3a g2 is a STU3N45K3, but most high voltage N chan will be fine there I just wanted one with a low Cgs

Posted: **Mon Feb 20, 2023 2:20 pm**

Nick wrote: ↑Mon Feb 20, 2023 1:29 pm What you have missed is the line that says that the depletion mosfets are IXTP08N100's I just used the 740 models as they have the same pinout. The mosfet that's driving the d3a g2 is a STU3N45K3, but most high voltage N chan will be fine there I just wanted one with a low Cgs

Ahhhhh... like footprints in KiCAD, yes I've used the same trick sometimes. That explains it.

Looks like I'll have to go and find a model for that, it's (predictably) not in the list of available NMOS devices...

Posted: **Mon Feb 20, 2023 2:48 pm**

Fixed it! Found a model on https://www.diyaudio.com/community/thre ... el.307318/ which makes use of the VDMOS model, pasted it into standard.mos and I was able to pick it as a transistor.

Sim seems to work now, I get about 44dB gain before it hits the RIAA and the various current figures seem to be about where they should be.

Posted: **Mon Feb 20, 2023 4:07 pm**

That sounds about right. In practice I get 40dB for the first stage, its controlled by the feedback percentage set by R22 and R23 in your model.

Posted: **Mon Feb 20, 2023 5:56 pm**

So post-RIAA I go a bit off-piste:

I was looking at the output stage of Stephen's (izzy-wizzy) circuit and couldn't help noticing a similarity to Lynn Olson's Raven. This is a version of that, with rough resistor values in there for now.

Gain of the whole thing is modelling at 40dB at the 1kHz point.

Posted: **Tue Feb 21, 2023 9:34 am**

With that lcr, don't you either have to drive it with a 600R source or terminate it with 600R?

I might have missed something with the driver as I'm looking at it on a phone.

Posted: **Tue Feb 21, 2023 9:45 am**

In the case of the original, its a 1k5 LCR and its terminated on the LCR board. The inductors are connected to a board that contains the caps and resistors.

Posted: **Tue Feb 21, 2023 6:38 pm**

Yes, of course, i asked you similar questions when I first saw the circuit.
So that's the reason for the 2 x 33uF caps from the 5687 CF - to get a low time constant with the 1.5K termination.

Posted: **Tue Feb 21, 2023 6:42 pm**

IslandPink wrote: ↑Sat Feb 18, 2023 5:01 pm ... the previous Pentode mu-stage, to help stabilise it ....

Sorry Gm stage

Posted: **Wed Feb 22, 2023 5:32 pm**

IslandPink wrote: ↑Tue Feb 21, 2023 6:38 pm Yes, of course, i asked you similar questions when I first saw the circuit.
So that's the reason for the 2 x 33uF caps from the 5687 CF - to get a low time constant with the 1.5K termination.

So with a 600R termination it looks like those values need to go up a bit.

Posted: **Wed Feb 22, 2023 5:54 pm**

The classic 600r lcr module can cause some issues which is why some of late have taken a different course of higher Z all the way to 10k. 1.5k or thereabouts is seen as a "better" bet to keep away the issues associated with much higher z.

Some get rid of the requirement of a huge coupling cap and DC couple to lcr and small cap after or something else. But then you need the lcr to not have a DC path to ground. I think that particular lcr configure isn't the most accurate configuration.

As always, there ain't no free lunch.

audio-talk

It's time I built a D3a -> 5687 phonostage

#16 Re: It's time I built a D3a -> 5687 phonostage

#17 Re: It's time I built a D3a -> 5687 phonostage

#18 Re: It's time I built a D3a -> 5687 phonostage

#19 Re: It's time I built a D3a -> 5687 phonostage

#20 Re: It's time I built a D3a -> 5687 phonostage

#21 Re: It's time I built a D3a -> 5687 phonostage

#22 Re: It's time I built a D3a -> 5687 phonostage

#23 Re: It's time I built a D3a -> 5687 phonostage

#24 Re: It's time I built a D3a -> 5687 phonostage

#25 Re: It's time I built a D3a -> 5687 phonostage

#26 Re: It's time I built a D3a -> 5687 phonostage

#27 Re: It's time I built a D3a -> 5687 phonostage

#28 Re: It's time I built a D3a -> 5687 phonostage

#29 Re: It's time I built a D3a -> 5687 phonostage

#30 Re: It's time I built a D3a -> 5687 phonostage