Booster v1.1 testing

[hartytp]

Booster v1.1 here and looking really nice. The mechanics feel vastly improved from the initial version.

One minor niggle is that it would be slightly nicer if the screws on the FP were CS rather than cap head. Otherwise, it looks great.

Also, the lid was a bit tricky to remove because it fits so snugly (had to stick a screwdriver through the fan holes to push it out). However, once I did get it off I was rewarded by seeing how beautiful it all looks inside. I'm really impressed.

[hartytp]

From a mechanical standpoint: I've done some quick disassembly (removed an amplifier module lid, removed an entire module) and I'm really struggling to find things I don't like about it. Very well done.

[hartytp]

Actually, one other thing: the (M3?) internal CSK Phillips screws were a bit tricky to remove. Not sure if they were threadlocked or just done up very tightly, but some of the heads started to strip. Using cap head screws (and maybe even increasing the screw size a touch) would probably make life easier.

Edit: or, for the ones that have to be CSK, can we at least use allen key heads?

[hartytp]

Also, channel 0 was a little tricky to remount. I had to remove the lid to get the D-type to remate with the PCB.

[hartytp]

Related to the previous comment, the amplifier mounting feels a little over-defined due to the use of CSK screws to hold the amplifier channels down. Might be better to use cap head screws with a generous clearance hole to give some slack.

[hartytp]

@gkasprow @wizath here is the UART output when I enable it:

• channel 1's interlock seems to trip straight away
• channel 2 has a very low bias current
• channel 6 has a very high bias current
• channel 7 has a low bias current

any ideas?

[hartytp]

Okay, the interlock on channel 1 was because I wasn't 50R terminating that input (none of the other channels seemed to mind).

[hartytp]

Still, the variance in bias current seems odd, since these have been calibrated.

[wizath]

Did you swap any module?

[hartytp]

I don't believe so! I wrote numbers on them before removing, so the worst case would presumably be if I had swapped 7 with 0, 6 with 1, etc. However, looking at the bias current measurements, I don't think that would explain what we see. Do you?

[wizath]

I've observed that calibration is channel specific - e.g if you calibrate module in slot 1 then insert it in slot 2 it won't be the same

[hartytp]

However, maybe we should make the UART report the unique ID of each channel, and store the channel ID along with the calibration data, to allow this kind of think to be checked.

[hartytp]

Okay, let me try swapping them around and see what happens...

[wizath]

I'll add ID reporting

[hartytp]

hmm...in fact, I can't have swapped the channels around. I can tell by the stickers that TechnoSytem put on them (the stickers bridge the gap beween adjacent amplifier channels, so if you swap the channels around the stickers aren't in the right places any more). So, I'm pretty confident that they are all in the same places that I received them in..

[hartytp]

See this photo. If I swap 7 with 0 then the stickers don't bridge the units correctly

PS these electro polished RF modules are so shiny I feel like I need to be in a clean room to touch them

[wizath]

Owwww, they could swapped them....

[hartytp]

Owwww, they could swapped them....

? how?

[hartytp]

Did you record the channel IDs when you calibrated them? that would allow us to check.

[wizath]

They took it from me to assembly them, and probably swapped them all. But its simple fix - I'll write you the details about current calibration since it's super easy

[hartytp]

Great, thanks! Can you post the details on the Wiki, please?

Do we also need to recalibrate the Tx/Rx power measurements?

[wizath]

No, I don't think power measurement recalibration is needed.

You can connect 50 Ohm terminator to input or leave it unconnected. Then you type stop in console to stop status messages from appearing (the're disabled by default in new version). Then you simply call calbias <ch number> <desired current> and wait if the process will catch the selected value.

[hartytp]

You can connect 50 Ohm terminator to input or leave it unconnected. Then you type stop in console to stop status messages from appearing (the're disabled by default in new version). Then you simply call calbias and wait if the process will catch the selected value.

Nice! is the bias current in A or mA? i.e. is the calibration process you did calbias 0 0.16?

I'm glad you implemented this nicely -- at some point I'd like to play with running these amplifiers at a lower bias current to reduce the power consumption at the expense of less gain and lower P1dB. This will make that process very easy

[wizath]

in mA's, calbias 0 160

[hartytp]

thanks

[wizath]

Anyway, I think that this option should be available to end-user since current calibration can't really be done properly in production.

[hartytp]

Anyway, I think that this option should be available to end-user since current calibration can't really be done properly in production.

Why not?

[hartytp]

@wizath for channel 2, calibration seems to have failed. I see DAC 15 CURRENT 120 Have you seen that before?

Anyway, I think that this option should be available to end-user since current calibration can't really be done properly in production.

If the calibration alogrithm were improved a bit (large steps first, then smaller steps) so that it was a bit faster then the amplifier could just autocal on power up...

[hartytp]

Channel 7 calibration also fails: DAC 15 CURRENT 146

[wizath]

Yes, I will improve it, but for now there are other issues in line.

Channel 7 calibration also fails: DAC 15 CURRENT 146

That's unfortunate - we need to better set the voltage, by using the potentiometer on module, so there is more range to calibrate on (since calibration values on some modules are very close to the DAC limits).

[wizath]

You can adjust the current (displayed on status messages), by using potentiometer on RFPA_PA_OUT board, then recalibrate.

[wizath]

Anyway, I think that this option should be available to end-user since current calibration can't really be done properly in production.

Why not?

If we calibrate whole set (box + 8 modules) it's fine, but if user wants to replace one unit, it must be calibrated again after inserting it to the slot - because AFAICT current calibration is slot and module dependant.

[hartytp]

That's unfortunate - we need to better set the voltage, by using the potentiometer on module, so there is more range to calibrate on (since calibration values on some modules are very close to the DAC limits).

Okay, maybe we need to adjust these limits for the next revision.

[hartytp]

You can adjust the current, by using potentiometer on RFPA_PA_OUT board, then recalibrate.

Where is that? Do you have a photo?

[wizath]

It's the POT1, in the middle

[hartytp]

Okay, I'll look at that. Any advice as to how I should tweak it (which direction, how much)?

[hartytp]

FWIW, in that channel, the coax was resting on the PA chip. I'll put some hot glue to hold it away.

[wizath]

Oh, I forgot - another command is bias <channel> <value 0-4095> which enable you to control bias voltage. You can set it to 0, then turn the potentiometer to lower the voltage (if there is not enough current). You can measure negative voltage on the pot's leg close to C2 silk. It should be around -0.8-1 V, then reapply calibration.

The current changes will be visible on status messages easily.

[wizath]

Next batch will be calibrated to the center of DAC range, so there will be more room to adjust the current.

[hartytp]

Oh, I forgot - another command is bias <value 0-4095> which enable you to control bias voltage. You can set it to 0, then turn the potentiometer to lower the voltage (if there is not enough current). You can measure negative voltage on the pot's leg close to C2 silk. It should be around -0.8-1 V, then reapply calibration.

I don't have a wiring harness to give me access to the pot while I do that. I could make something but it would take time

[wizath]

Oh, I forgot about top board blocking the way. I'll tell you which way to turn to increase current in a second

[wizath]

Maciej suggested that top board should have hole for screwdriver to adjust the pot.

[hartytp]

Maciej suggested that top board should have hole for screwdriver to adjust the pot.

:+1: and a note on the silk screen indicating which direction to turn for more juice

we should also consider changing the dividers to give a large range of bias currents without a pot. Or, even scrap the pot (do we need it when the DAC is 16-bit?)

[wizath]

So you turn in direction of "+" to get more current, and opposite to get less current. Be gentle, and turn only several degrees. DAC is 12 bit

[hartytp]

DAC is 12 bit

Would it be easier to use a higher-resolution DAC in the next revision and scrap the pot? One fewer thing that has to be adjusted by hand.

[hartytp]

Thanks. That works

[hartytp]

Any idea why channel 2 interlock trips if the input isn't 50R terminated? No other channels do that.

[hartytp]

[wizath]

Any idea why channel 2 interlock trips if the input isn't 50R terminated? No other channels do that.

That may be some issue with power detector. I've got module with simmilar issue - waiting for Maciej to come back from Germany and resolve this.

[hartytp]

Okay, thanks.

One more thing. Look at this:

why is channel 7 so hot?

[hartytp]

hmm...unless it's the hot glue I put in to hold the cable down keeping the sensor warm