Cubie3

Thanks again juma, will report back.

Yes, the main pcb will have the circuit for the meters and ssr speaker protection.

Edit: and the connector for the headphones.
 

Attachments

  • Screenshot_9.jpg
    Screenshot_9.jpg
    73 KB · Views: 107
  • Like
Reactions: von Ah
...DC offset is showing around 71mv after 30min of working ,is it normal ?...
DC offset up to 100mV is not a problem but if you want to make it lower, the easiest way to do it is to put a serial connection of 120R resistor+ 50R multiturn pot instead of R8 or R5 so you can adjust it.

Of course, there are other methods: transistors' thermal coupling and/or transistors' matching as well as adding a DC servo circuit etc... There are a lot of posts/threads on these subjects, be free to search the forum/net. :cheers:
 
Last edited:
  • Like
Reactions: VishnuS
Since there are several solutions, it would be easier to offer the easiest solution if you would include more information of your layout, especially a photo. You will notice from post #1 that juma included a pot in series with a slightly smaller value of R8. If adding in a pot makes it too messy on your layout, then you could find it easy to parallel a resistor with R5 to bring the R8 to R5 leg slightly more negative. That is if the present offset is +71mV and not -71mV. Maybe try a resistor value between 2k2 and 3k3. Small steps are better than big jumps. After you see the amount of change you can use the formula for parallel resistance to calculate a better value resistor.
Would you also measure and post the amount of heat rise you have along with a photo of your heat sink? That may help others judge how much heat sink they need for this circuit. It looks like a very nice circuit that needs only low-cost, easy to obtain parts.
 
  • Like
Reactions: VishnuS
DC offset up to 100mV is not a problem but if you want to make it lower, the easiest way to do it is to put a serial connection of 120R resistor+ 50R multiturn pot instead of R8 or R5 so you can adjust it.

Of course, there are other methods: transistors' thermal coupling and/or transistors' matching as well as adding a DC servo circuit etc... There are a lot of posts/threads on these subjects, be free to search the forum/net. :cheers:
Thanks sir
 
Since there are several solutions, it would be easier to offer the easiest solution if you would include more information of your layout, especially a photo. You will notice from post #1 that juma included a pot in series with a slightly smaller value of R8. If adding in a pot makes it too messy on your layout, then you could find it easy to parallel a resistor with R5 to bring the R8 to R5 leg slightly more negative. That is if the present offset is +71mV and not -71mV. Maybe try a resistor value between 2k2 and 3k3. Small steps are better than big jumps. After you see the amount of change you can use the formula for parallel resistance to calculate a better value resistor.
Would you also measure and post the amount of heat rise you have along with a photo of your heat sink? That may help others judge how much heat sink they need for this circuit. It looks like a very nice circuit that needs only low-cost, easy to obtain parts.
Currently, I don’t have anything to measure the temperature. After 3 hours of working, temperature of lateral mosfet is normal only.

I think below mentioned heatsink size is appropriate

Length-25cm Width-4cm Height-11.5cm


Please find the attached snap for reference
 

Attachments

  • PCB front.jpg
    PCB front.jpg
    375.7 KB · Views: 104
  • PCB back.jpg
    PCB back.jpg
    342.4 KB · Views: 107
  • Silkscreen.jpg
    Silkscreen.jpg
    106 KB · Views: 116
  • Copper side.jpg
    Copper side.jpg
    84.1 KB · Views: 107
  • Heatsink.jpg
    Heatsink.jpg
    251.4 KB · Views: 107
  • PCB + Heatsink.jpg
    PCB + Heatsink.jpg
    381.3 KB · Views: 100
  • Heatsink 1.jpg
    Heatsink 1.jpg
    263.4 KB · Views: 93
  • Heatsink 2.jpg
    Heatsink 2.jpg
    254.4 KB · Views: 102
DC offset up to 100mV is not a problem but if you want to make it lower, the easiest way to do it is to put a serial connection of 120R resistor+ 50R multiturn pot instead of R8 or R5 so you can adjust it.

Of course, there are other methods: transistors' thermal coupling and/or transistors' matching as well as adding a DC servo circuit etc... There are a lot of posts/threads on these subjects, be free to search the forum/net. :cheers:
Sir, can i use 100ohm multiturn port?
 
Completed the cubie3 JFET version with k246 & j103 JFETs

Sound is amazing 😱
 

Attachments

  • IMG_20240423_205356.jpg
    IMG_20240423_205356.jpg
    476.8 KB · Views: 94
  • IMG_20240423_205406.jpg
    IMG_20240423_205406.jpg
    358.8 KB · Views: 91
  • IMG_20240423_205452.jpg
    IMG_20240423_205452.jpg
    322.8 KB · Views: 85
  • IMG_20240423_205429.jpg
    IMG_20240423_205429.jpg
    331.4 KB · Views: 77
  • IMG_20240423_205442.jpg
    IMG_20240423_205442.jpg
    235 KB · Views: 79
  • IMG_20240423_205412.jpg
    IMG_20240423_205412.jpg
    376.4 KB · Views: 83
These days I finished my Cubie.
It uses 7ma jfets, j77 and k216 in the vas, 08n16v and 08p16v outputs.
IMG_7611.jpeg


I run the vas at 30ma.

ee1f572a-8960-41df-a189-1df7bf71f195.jpeg


When I increase the bias above 100ma the amp starts to oscillate.
If the input is not connected the oscillation is smaller
IMG_7616.jpeg


If I ac couple the input to a signal source I get this
IMG_7615.jpeg


Power comes from a 28vac dual secondary trafo

What could be the cause of this?
 
Seems I spoken too soon.

Yesterday I removed the gate resistors from the vas and set the bias fixed at 130ma for the each channel by replacing the pot with a fixed resistor.
Ran a couple of square waves and everything looked fine.
The front end runs at 5.5ma, the vas at 20ma and no sign of oscillation.

Today I placed back the pots to be able to set a higher bias and the oscillation appeared back.
When I increase the bias in the ops the mosfets in the vas go south and draw 50ma.

Any ideas what can I search for?