Referenz - high-quality low-cost diy-friendly amplifier

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
The design files were updated on January 28, 2019 - please use V1.3 files from post #103

The time has come to publish something really DIY-friendly ;)

Please welcome Referenz - a relatively simple low-cost amplifier, utilizing current-mode amplification.
Low distortion, low noise. Power rating: 100W @ 8 ohm / 200W @ 4 ohm.
The amplifier consists of IPS module and OPS module.
Two OPS variants are included, however, the layout for quasi variant is still under development.
OPS employs commonly-available inexpensive HexFETs.

The amplifier is live-tested - please see the test results in the post #97.

Jeff has designed cool single-sided layouts, slightly tuned-up based on the tests of my prototype.
Questions are welcome - the builds will be supported.

Cheers,
Valery
 

Attachments

  • 01 Referenz Input @Sch 1.2.jpg
    01 Referenz Input @Sch 1.2.jpg
    166.8 KB · Views: 1,985
  • Referenz Files.zip
    357.5 KB · Views: 169
  • @Referenz Loop Analysis.jpg
    @Referenz Loop Analysis.jpg
    132.4 KB · Views: 530
  • 03 Referenz Output Quazi.JPG
    03 Referenz Output Quazi.JPG
    96.7 KB · Views: 555
  • 02 Referenz Output.JPG
    02 Referenz Output.JPG
    98.5 KB · Views: 1,741
  • 01 Referenz Input.JPG
    01 Referenz Input.JPG
    66.6 KB · Views: 1,842
  • 03 Referenz Output Quazi @Sch 1.2.jpg
    03 Referenz Output Quazi @Sch 1.2.jpg
    210.4 KB · Views: 1,923
  • 02 Referenz Output @Sch 1.2.jpg
    02 Referenz Output @Sch 1.2.jpg
    209.9 KB · Views: 2,005
Last edited:
The time has come to publish something really DIY-friendly ;)

Please welcome Referenz - a relatively simple low-cost amplifier, utilizing current-mode amplification.
Low distortion, low noise. Power rating: 100W @ 8 ohm / 200W @ 4 ohm.
The amplifier consists of IPS module and OPS module.
Two OPS variants are included - complementary and quasi-complementary.
Both variants employ commonly-available inexpensive HexFETs.

It's not live-tested yet, so early adopters may face some amendments.
At the same time, some solutions used in this design are rather mature, being used earlier in the other designs.

Jeff has designed cool single-sided layouts.
Questions are welcome - the builds will be supported.

Cheers,
Valery


looks good , like my first CFA.


One comment , you chose to servo the low Z FB.
Have you tried either servo(ing) the CCS's or the high Z input.
Thimios did all 3. The high Z worked , 1M ohm Z load for the op-amp.
Servo FB to the CCS's also worked , but added more components.


PS- with especially an extreme input pair mismatch , the servo does
not have the "horsepower" to compensate the current FB node.
Servo FB is the "anal" way , since FB is independent of the audio paths.


OS
 
Hi Pete!
Glad to see you here!
Driled, not my best but still usable, i hope.
If changes, i hope some flying parts will do the job. :confused:
This time nothing works first time. :(
I haven't any photoresist remaining and the transperance not function well. :(
 

Attachments

  • IMG_20190107_185247.jpg
    IMG_20190107_185247.jpg
    722.2 KB · Views: 816
looks good , like my first CFA.


One comment , you chose to servo the low Z FB.
Have you tried either servo(ing) the CCS's or the high Z input.
Thimios did all 3. The high Z worked , 1M ohm Z load for the op-amp.
Servo FB to the CCS's also worked , but added more components.


PS- with especially an extreme input pair mismatch , the servo does
not have the "horsepower" to compensate the current FB node.
Servo FB is the "anal" way , since FB is independent of the audio paths.


OS

Hi Pete,

Thank you for the observation - the servo connection is actuallu a mistake :)
It happened when redrawing the schematic in Diptrace.
Jeff - we've got a problem :p

56K resistor has to be connected not to the low Z FB node, but to the high Z input (input pair's bases). This is an inverting servo, so it is intended for controlling the input, opposite the one with the feedback.

Thimios - sorry, you will need to cut the trace and reconnect the 56K resistor's lead.

Pete - I thought about driving the CCSs with the servo, but this design had a special requirement, driving my choices - it had to be as low-cost and low part count as possible (still showing very good sound quality characteristics).
Driving the high Z input with servo is a proven solution I used in the past. 56K resistor's value is chosen for having the overall input Z at the level of 27K (in parallel with 47K connected to ground).

We will upload the corrected files to the Post #1 soon ;)

Cheers,
Valery
 
UPDATED FILES

Here are the corrected files with all amendments implemented and etching-friendly pads on all the boards.
I will update post #1 as well, however, the system doesn't allow me updating the files there (only allowing updating the text).
 

Attachments

  • @Referenz Loop Analysis.jpg
    @Referenz Loop Analysis.jpg
    132.4 KB · Views: 195
  • 03 Referenz Output Quazi.JPG
    03 Referenz Output Quazi.JPG
    96.7 KB · Views: 177
  • 02 Referenz Output.JPG
    02 Referenz Output.JPG
    441.8 KB · Views: 189
  • 01 Referenz Input.JPG
    01 Referenz Input.JPG
    66.6 KB · Views: 307
  • 03 Referenz Output Quazi Sch 1.2.jpg
    03 Referenz Output Quazi Sch 1.2.jpg
    154.8 KB · Views: 689
  • 02 Referenz Output Sch 1.2.jpg
    02 Referenz Output Sch 1.2.jpg
    149.4 KB · Views: 705
  • 01 Referenz Input Sch 1.2.jpg
    01 Referenz Input Sch 1.2.jpg
    129.8 KB · Views: 708
  • Referenz Files 01 Corrected PDF.zip
    349.7 KB · Views: 130
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.