| wai235 |
I've not built the original Pearl. I only have a MC cartridge (AT OC9MLII). I want the Pearl has enough gain for my OC9. Thus, I tried to modify the circuit for more gain. The modification was made with following considerations:
- Increase the total gain to over 60db.
- Maintain the two amplification stages of the original Pearl design. I think addition of one or two more amplification stages to the original Pearl design will increase noise and distortion.
- Minimum of components (it is the objective of original Pearl design).
Details of modification:
1. Increase the number of 2SK170BL of the first stage amplification to 8.
2. Reduce the resistance of R28-R31 to 18R.
3. Increase the resistance of R19 to 619R.
4. Replace the 2SK389 of the second amplification by 6 2SK170BL.(going to acquire 100 2SK170BL for 2x8 and 2x6 matched 2SK170BL.)
5. Increase the resistance of R14 to 1K5.
6. Use two 12V SLA batteries as power supply. The entire circuit of power supply regulation can be discarded. Two fully charged 12V SLA batteries can provide 30V. I think it should be sufficient and can provide the similar voltage as stated in original circuit.
The modification seems to be simple. I don't know whether I overlooked something. Would anyone tell me whether the circuit work or not?
Thanks
Wai |
|
|
| wai235 |
Hi all,
I just got 200 2sk170bl. I going to measure the Idss first. Just thinking what Idss value should I use (high or low).
:confused: |
|
|
| wai235 |
Hi All,
All the 2sk170bl come with high Idss. I could only get 14 of them with Idss 13.5mA. 9 of them with 13.6mA. The Idss was measured with Vds 9Volt. The datasheet states that Idss of 2sk170bl is between 6mA and 12mA. All the 2sk170bl I got are over 13mA. :bawling: :bawling:
Do I need to get 2sk170gr?
Wai |
|
|
| wai235 |
I decided to amend my original modification:
1. 2SK170GR will be used instead of 2SK170BL. It is because the drain current of 2SK170BL is too high for SLA batteries.
2. In order to maintain sufficient Vds for the 2SK170, 3 x 12V SLA batteries will be used.
3. Change the R19 back to original value 499R of the original Pearl design. R19 determine the output impedance of the first stage amplifier. Any change on this R19 will upset the RIAA network and the accuracy of the RIAA network cannot hold.
4. Change the R14 back to original value 1k of the original Pearl design. I think increase the value the value R14 is not good for low input impedance pre-amp.
5. The decrease of values of R19 and R14 will decrease the gain. This can be compensated by reducing the value of the source resistors of the first and second stage amplifier to 7R8. This value came up with assumption of 2SK170GR (Idss=6.5mA) is used. I also don’t want the value to be too small for any chance of forward bias the 2SK170.
To sum up, the additional gain of Pearl will be obtained by two modifications only. They are the parallel of more 2SK170 and the decrease of the value of the source resistors.
Any opinion is welcome. :worship:
Wai |
|
|
| wai235 |
Hi all,
I got another bag of 200 2SK170GR. :D I measured all Idss. Most of them are 4.6 - 5.1mA. I have to re-calculate the configuration because I assume a 6.5mA in my previous calculation. Since they are low transconductance, I have to parallel more 2SK170GR in each amplification stage.
To be an audiophile :cool: , I am going to measured Vp the pinch-off voltage in order to get the transconductance values. I'll parallel 2SK170GR with similar transconductance in order to reduce noise.
Wai |
|
|
| WALTER BURKHARD |
Hi wai235 can u contact me???
regards walter |
|
|
| wai235 |
Hi Walter,
Sorry for my late reply.
My diyAudio account is still under moderation. I can't reply to you through diyAudio mail system. Could you send email to me at pcwcheng@gmail.com ?
Regards,
wai235 |
|
|
| wai235 |
Hi everybody. Although I've not reported the progress for long time, I've not given up the project.
I've tried to do simulation in order to verify the circuit. Thus I get the free LTspice/SwitcherCAD III http://tech.groups.yahoo.com/group/LTspice/
After simulation with LTspice, I give up the two-stages amplification approach with the following reasons:
- the gain of the first stage is too high and will forward bias the JFET of the second stage
- the Vce across those ZTX450 is too low (this may lead to unstable Ice current and distortion) even with 3 x 12V SLA batteries as power supply
- can't achieve net gain 60db
I tried to modify the circuit with a three-stages amplification approach:
- maintain a low Vds of all JFET (around 8V) to minimize the gate leakage current (ref: http://www.klaus-boening.de/html/lab_page.html )
- maintain sufficient Vce across all ZTX450
- although three-stages amplification will make the output signal out of phase with input, I don't want to add a fourth amplification stage for just correct the phase
- since a three-stages approach allow me to split the RIAA filter to two filter sections, I decide to include the fourth time constant (ref: http://www.klaus-boening.de/html/ti...time%20constant )
New circuit is attached. http://i46.photobucket.com/albums/f...235/pearlv4.png
Comment is welcome.
Thanks
Wai |
|
|
| wai235 |
| With the inverse RIAA as input, the attached plot shows that the accuracy of the RIAA filter is within 0.2db from 10Hz to 100kHz! |
|
|
| wai235 |
| This plot shows the new circuit can achieve a gain of 60db. :) |
|
|
| wai235 |
| This plot shows the input (blue line) and the output (green line). |
|
|
| Panelhead |
I built mine with an integral pre- preamp stage to allow use of LOMC cartridges. Tried a couple nice MC step up transformers and neither seem to drive the Pearl very well. Suspect the gate capacitance of the four 2SK170BL was too much.
Copied the input stage of the Pearl to achieve this. Just used two 2SL170V as the stage. Used the cascodes, cap coupled. Also pulled power from last discrete reg and used an LC filter on the added stage.
With almost no changes to the basic Pearl this is clean, fast, and quiet. Much better than the transformers.
I think your attempt at modifing the two stage would have netted minimal gain increase, maybe 6 dB at best. Slapping the activestage in should net another 25dB in practice, so it boosts the Pearl from 44 dB up to almost 70 dB.
George |
|
|
| wai235 |
Hi Panelhead,
Thanks for sharing your experience. :worship:
| quote: | | Copied the input stage of the Pearl to achieve this. Just used two 2SL170V as the stage. |
Why did you use 2SK170V instead of 2SK170BL?
| quote: | | Also pulled power from last discrete reg and used an LC filter on the added stage. |
Could you show me the details of the LC filter?:worship:
| quote: | | I think your attempt at modifing the two stage would have netted minimal gain increase, maybe 6 dB at best. Slapping the activestage in should net another 25dB in practice, so it boosts the Pearl from 44 dB up to almost 70 dB. |
Yes, I give up the two-stage approach.
Thanks
Peter |
|
|
| wai235 |
Hi all, here are some photos of my new phono amp.
Since I only got two small metal cases, I built two mono blocks! :cool: I finially use 2 SLA batteries for each channel.
Wai |
|
|
| vinylkid58 |
| quote: | Originally posted by Panelhead
Tried a couple nice MC step up transformers and neither seem to drive the Pearl very well. Suspect the gate capacitance of the four 2SK170BL was too much.
George |
That's very interesting. I was thinking of going the SUT route for my Pearl, but maybe I won't bother now based on that tidbit of info. You maybe the first to mention this.
Thanks,
Jeff |
|
|
| wai235 |
Hi all,
I used this circuit layout for the soldering. I designed the layout with the following objectives:
1. star grounding
2. floating the entire circuit from the chassis
3. use component lead only (except grounding wire), otherwise use shielded cable for signal path
4. use the row of 1000uF capacitors as shielding for separating the 3 amplification stages
The result is good. No hum! :cool:
Wai |
|
|
| KBK |
I would suggest some polystyrene or better bypass caps on the big Solen cap, this will create great sonic improvement.
My kind of circuit board in there! IE, none! Next time, consider using fairly large blobs of glue to lock the components down, in order to distance the components even further away from the metal chassis. Inductive reactance from the chassis slurs the phase at HF frequencies, giving false highs and false detail. |
|
|
| wai235 |
| quote: | | I would suggest some polystyrene or better bypass caps on the big Solen cap, this will create great sonic improvement. |
Hi KBK,
The bypass cap is a 10uF solen ( Metallized Polypropylene Film Capacitor). I think it is good enough. However, I won't refuse to make the Pearl better if the cost is not too high. Please name suggestion cap with brand name and model.
Thanks
Wai |
|
|
| wai235 |
| quote: | | consider using fairly large blobs of glue to lock the components down, in order to distance the components even further away from the metal chassis. |
Entire circuit is suspended in the air. Sound great. Just wonder how to do it. Could you show me picture?
Thanks
Wai |
|
|
| wai235 |
Hi all,
I'm still enjoying my Pearl. Just wake up and thinking I should post the spice model here. The file includes the measured value of all the capacitors used in the RIAA filter. The model let me see the performance of the riaa filter. I used this model to select the suitable resistor values.
This circuit is highly customized for my need (say the AT OC9 cartridge) and the components I could get. You may want to adjust the value of the components to suit your need. Thus, I post the spice model here. Please also note that I ajust the RIAA filter section such that the low start to roll off at 8Hz which is much lower than the original Pearl. If you encounter rumble problem, I suggest you decrease the value of C13.
I also attached the modified standard.bjt and standard.jft files of LTspice which include spice model of 2SK170GR and ZTX450. You can use them to replace the original files stored at c:\Program Files\LTC\SwCADIII\lib\cmp for running my spice model of Pearl.
Regards,
Wai |
|
|
|
