Ultrasimple mm/mc RIAA preamp 2

[jonners]

2SK170BL is still available from Spencer through this forum: Toshiba 2SK170BL for sales

[ra7]

Hey Salas,

Like I mentioned before, I learnt a little spice and ran some simulations on the original pacific and the MadK version.
Both the version show severe clipping at 20 Hz. Of course, I'm at a very early stage learning spice. So, I may be wrong.

I built the one you gave me and listened to it, but it wasn't sounding all that good. Part of it, I thought was due to the high source resistance necessitated by my 8mA Idss JFETs. I went ahead and started playing with it in spice - changing the load resistor values, power supply voltage and the source degeneration resistor. Optimizing for gain, lowest distortion, and lowest feedback (of the degenerative kind), trying for that "sweet spot."

I ended up with the attached circuit. Reduced the supply to 16V. It gives me enough gain, no clipping at 20 Hz. I built this one as well and it sounds wonderful. I think it sounds better than the original, but of course I'm biased.

It shows about 0.4% distortion at 1kHz, which is high by hi-fi standards. But this is the best I could conjure. There are a lot of threads for this circuit, MadK has a link on the first page of this thread, and the next possible improvement is a cascode BJT or JFET on the top of the two gain devices. But that is beyond my understanding right now.

Feel free to suggest improvements or if I am doing something incorrectly.

[ra7]

One thing I didn't understand was how the original circuit which was clipping badly at even 50 Hz was considered at all. When I first listened to it of course it sounded great, but it was distorting on peaks and on complex passages.

We are supposedly more tolerant of distortion in the low frequencies than high. But all the phono preamps spec the distortion at 1k. In an application where the gain at 20 Hz is about 20 dB higher than at 1k, isn't what is happening down there important?

Am I missing something?

[Salas]

Its your cartridge proved being too strong for them circuits. ADC 8mV! You did well to optimize with it. Good work.

P.S. One improvement would be to decouple the first stage B+ with RC 100R 1000uF (R from B+ to drain, C from drain to gnd). There is interaction. And also to make some good reg for it at a point.
R2 must be 1k5 too on your schematic, yes? 2k2 does not hold any Vds to speak of with 16V B+ 8mA Idss and 5R Rs. Lets check, about 5.5V under R8 you get?
You kept the better Riaa values, good. Did you trim your RinLoad by listening and tweaking a 100k trimmer at all?

[ra7]

Salas,

R2 is 2k2 and I'm using 2k2 in the built version. I checked the Vds, it was about 4V on the sim, I will check it today in the actual circuit and tell u. I will try it with the 1k5 resistor.

Why is important to setup Vds high?

I did not try the 100k input trimmer. I will do that as well. What are we looking to do here? I mean what is going to cause the sound to change?

Also, any pointers on what can be done to improve distortion from here?

Salas, you've been a great inspiration. Thanks! And thanks to the other folks here for helping me find other tidbits of information on the sim.

[ra7]

I was looking at the pearl phono on passdiy and just realized that it is essentially a highly embellished pacific phono preamp.

At the input 4 JFETs are paralled and cascoded with small source degeneration. Then the eq is applied between the first and second stage. Then again the second stage is paralled and cascoded.

But the framework of the circuit and the gain devices used are the same.

hmm... something to ponder, ain't it?

[Salas]

Because of low degeneration 5R, just 16V B+ and high Idss of the BL you use compared to the GR that the circuit you started with used, I am expecting too low Vds in the first stage. Unless your 2SKS are less than 8mA? Did you check them off circuit with a 9V battery, having G&S tied together and a DVM in current mode in series? Its about having some headroom since there it is pre Riaa and you got an 8mV MM cart. If you really got 4V under 2k2 is rather OK. But see that you haven't got far low by a simple measurement.

On MM carts, varying the load resistor from the typical 47k that nobody varies, has some interest. Look at a Werner article for TNT. Werner is a member in DIYA too.

Don't be thinking much about buzzwords like Pacific etc. Those are just titles to give for a certain effort, it was on a French Mag, had to have a title. Does not mean that successive common source stages aren't something very common logic to work with, or that the Pearl had anything needing to lend from the Pacific. Certainly not its whimsical Riaa values. Pearl has a low source resistor but 4X the current. I mean think about it, many DIY tube phono stages are successive common cathode with a passive filter in between. They are older than anything else, its just that nobody thought of naming one as ''The Atlantic''.

[ra7]

I got what you are saying and I agree.

Measured the drain voltage, about 5.5 Vds on the first stage.

I read the article. Very interesting. I will try it.

For batteries, I'm using two 9.6V rechargeable Ni-mh batteries, 1600mAh each. They drop 3-4V every hour of operation. The total current draw is about 22mA. Is there any fix to this?

[Salas]

5.5V with 2k2 RL and 8mA idss JFET, 5R Rs and 16V B+? Maybe you had stronger battery at the time? Are your 2SKs actually BL? Can you post your 2SK170 sim model?

To skip the battery pain just make a reg.

[ra7]

Yes, I think I had about 18V at that time and also in the circuit I have 10R Rs, changed from what I posted here. Sorry, my mistake.

Here is the model I'm using:
.model 2sk170 NJF(VTO=-.45 BETA=40M LAMBDA=19.95M RD=3.85 RS=2.35 IS=3.95F PB=1 FC=.5 CGS=85.4P CGD=22.7P)

I have about 4.55V in the sim right now.

Adjusted beta and Vto for Idss of 8mA.

The JFETs were from spencer who is a member here. They were in quads matched at 8.1 mA Idss.