A differential output PEARL... ...?

[xavier1000]

Analogair,

They are the same project. The MC stage I am using for my Shelter 910 cartridge is the one described in the other thread, which is very similar to the input stage of the Pearl. Additionally, I needed a low noise negative rail for the differential output stage. This psu is exactly the same as the one for the Pearl, but inverted.

So, why not to use an additional Pearl PCB for this input MC stage on one hand and the negative rail on the other (with all the circuits inverted)? This way, more than 75% of this additional Pearl pcb could be re-used.

You will find below a picture of the circuits I am testing. The upper channel is just the MC input stage + an standard Pearl. The channel below has the MC input stage and the negative rail in the left pcb, a Pearl input stage + the RIAA filter pcb and an additional perf board on top, with the differential output stage.

Xavier

 

[xavier1000]

PCB Design: HELP

This is the latest version of the PCB with some minor enhancements.

I plan to order a prototype of this PCB next week, but since it is my first pcb design I would need some help. Could anyone take a look and give some feedback?

Thanks!
Xavier

 

[rpafenberg]

Xavier,

How did your prototype PCB go?

I am interested in doing something like that myself.

Roy

 

[xavier1000]

Hi Roy,

The prototype PCB was OK. I did some improvements and tunning on the circuit during the last months:

- Removed the capacitance multiplier Q6 and live just C15a (3300uF) in order to increase the voltage at the collector of the two cascodes (now 17v)
- The value of R38 needs to be lower (I don't recall the exact value now)
- Also added two caps from the base of the cascodes to gnd (56uF)
- I would recommend to use a trimpot instead of R35 to be able to adjust the bias.

My main objective with this circuit was to have a Pearl with balanced output but without loosing the character of the Pearl. So I left one of the channels untouched and used this circuit in the other channel to be able to compare.

I did all the tests in SE and I couldn't hear any difference when listening music. The only difference I could hear was some slight white noise only noticeable when cranking the volume at max (I usually listen at volume level lower than 1/3).

I believe this white noise comes from the CCS. In fact I've been testing several alternatives such as using zeners instead of leds or increase the voltage to allow increasing R35. I even tryed using a resistor instead of the active CCS, but the negative rail was not high enough. The less noisy solution that I tested was the two red leds in paralledl with the cap bypassed by 47n cap.

I destroyed one of the jfets in one of the tests. To fix this implies to unsolder both K389 (7 pins each!) and replace them with another matched pair. This is pending.

This is the current status.

I hope it helps.

Xavier

 

[rpafenberg]

Xavier,

Thanks for responding!

Glad to see you are making progress.

Looking at your schematic, I notice that c12a is listed as 10uF. Because both C12a and R24a are not on the PCB I assume that they refer to to C12 & R24 on the Pearl PCB and that the 10uF value is a misprint. Am I right?

Also the schematic lists Q4a & Q4b as ZTX450s while the PCB lists them as BC547Cs. I assume that the schematic is correct?

Regards,

Roy

 

[xavier1000]

Because both C12a and R24a are not on the PCB I assume that they refer to to C12 & R24 on the Pearl PCB and that the 10uF value is a misprint. Am I right?

Yes you are right, you can use C12 & R24 on the Pearl.

Also the schematic lists Q4a & Q4b as ZTX450s while the PCB lists them as BC547Cs

I tried both and, although both worked fine, I would stay with ZTX450 since seems to be less stressed.

Don't forget to double-check the pin position of the transistors in the PCB since ztx450 and BC547 don't use the same pinout.

Xavier