F5X -- the EUVL Approach - The Build Thread

You can check the accuracy by putting in ice water and then in boiling water, as Scott suggested.
You should get somewhere close to 0°C and 100°C.

Google triple point of water, even at home you can get better than .1C with a little care. It is considered the reference standard for thermometry and NIST has instructions for getting .001C. We did all this to qualify AD590's for the space shuttle. Review the history of the heat shield on the shuttle and you will get a harsh lesson on the value of obsessive measurement.
 
Do I need to connect the input's gnd to the chassis gnd? Also, since my bias is around 0.230v, should I reduce the fuse rating to 4amp slow blow as well? I'm about to close up the chassis.

T4A on a 220Vac mains supply allows a 900VA transformer to be used. What size of transformer did you buy?

Spec was 2x 600VA............
If that is what Pmc bought then Pmc needs two T3.1A mains fuses. Fuse each transformer separately. You can choose to have a single mains fuse earlier in the power circuit and the single common fuse should be around T6A, or T8A.
 
It even bend the jumpers from the input board to the main amp board. Any clue to which part that I should start investigating?
 

Attachments

  • 20170414_234213.jpg
    20170414_234213.jpg
    397.2 KB · Views: 390
I connected the gnd from the input board to the chassis gnd. Somehow, it is connected to the (-) output of the amp. Is this so? I thought it was supposed to be floating.

Input ground > amp ground > disconnecting network or CL-60 or similar > chassis/earth ground.

(-) and (+) outputs might be connected to amp ground via 220Ω (R21 & R22), if you chose to use it.
 

Attachments

  • F5X PCB front 1.jpg
    F5X PCB front 1.jpg
    446.5 KB · Views: 375
Last edited:
> Yes, I used the R21 and R22 resistors.

No, they have nothing to do with your problem.
Everyone, including me, used them and we have no problems.

> Should I remove these or should I just remove the CL-60 thermistors?

None of the above.
You have a short of your output to Gnd, according to you.
I would first check that, e.g. whether the Drains of your MOSFETs are touching the heatsink electrically.

> I just noticed there are 2 smd capacitors between the trim pots.

Optional, and you should not have soldered the electrolytic cap if you are not using cascode, as shown in your photo.

I suggest you do your testing with lab power supplies with current limit.
And you should short the inputs to Gnd during testing.


Patrick