Booster: metastability - Githubissues

hartytp commented 6 years ago

@gkasprow something concerning happened during my gain measurements

I plugged my VNA into one channel of booster and measured the gain. It was a bit lower than I expected
After a minute or so, I pressed the standby button to turn all channels off, then pressed it again to turn them on
after the channel came back up, it had the gain I expected (see plot)
I haven't been able to reproduce this issue. Any ideas?

metastability

Edit: fixed plot

gkasprow commented 6 years ago

Didn't you damage your attenuator in the meantime ? :)

hartytp commented 6 years ago

Didn't you damage your attenuator in the meantime ? :)

No. It's a 30W attenuator, with 100mW going into it. Moreover, the change in output power happened without me touching my measurement apparatus. I then remeasured the gain on other channels and found the same results I'd measured previously.

gkasprow commented 6 years ago

Please switch it off, let it cool and repeat. It looks like poor soldering in the gain-matching attenuator since the effect does not depend on frequency

hartytp commented 6 years ago

You think it's that and not some issue with the bias current starting up?

But, yes, I'll try that.

gkasprow commented 6 years ago

@wizath do you think this could be firmware issue? I'm not sure if gain changes that much with current. @hartytp you can connect the console and observe how drain current changes during measurement. At least if this happens again you'd have a history what happened with currents. Of course the current depends on the output power but then you can increase the power to get same output power and then compare the drain current.

wizath commented 6 years ago

DAC voltages are consistent through channel power cycle. I've observed this issue some days ago - even after calibrating the current, it sometimes differs from the set value. We can double check if the negative voltage is set properly during channel init.

hartytp commented 6 years ago

I've observed this issue some days ago - even after calibrating the current, it sometimes differs from the set value.

Do you mean that you observed the gain sometimes being different when the channel starts up (i.e. the same thing I observed)?

wizath commented 6 years ago

Not really the gain, just the current measurements were a bit off sometimes on single channel.

hartytp commented 6 years ago

Did this happen on all channels, or was there just one dodgy channel?

How much was the current off by? Enough to explain the gain changes I've seen?

gkasprow commented 6 years ago

@wizath was it the same channel?

wizath commented 6 years ago

Which channel number is bad? I think it was 4th

hartytp commented 6 years ago

Channel 5 was the problem

hartytp commented 6 years ago

Here is the diagnostic output at startup

PGOOD: 1
FAN SPEED: 0 %
AVG TEMP: 27.52 MAX:30.00
CHANNELS INFO
==============================================================================
        #0  #1  #2  #3  #4  #5  #6  #7
DETECTED    1   1   1   1   1   1   1   1   
TXPWR [V]   0.04    0.12    0.05    0.05    0.06    0.05    0.07    0.05    
RFLPWR [V]  0.08    0.04    0.05    0.04    0.04    0.09    0.05    0.04    
TXPWR [dB]  -12.02  -11.89  -13.18  -14.15  -12.34  -14.00  -13.02  -12.53  
RFLPWR [dB] -14.77  -11.67  -15.08  -15.10  -11.52  -14.66  -13.35  -13.66  
P30V [A]    0.118   0.159   0.164   0.144   0.169   0.134   0.162   0.167   
P6V0 [A]    0.024   0.032   0.033   0.029   0.034   0.027   0.032   0.033   
PN8V0 [mA]  0.501   0.681   0.696   0.613   0.714   0.572   0.688   0.712   
ON      1   1   1   1   1   1   1   1   
SON     1   1   1   1   1   1   1   1   
IINT        0   0   0   0   0   0   0   0   
OINT        0   0   0   0   0   0   0   0   
SINT        0   0   0   0   0   0   0   0   
SINTE       0   0   0   0   0   0   0   0   
SINTV       0   0   0   0   0   0   0   0   
OVC     0   0   0   0   0   0   0   0   
ADC1        70  195 82  82  99  86  113 83  
ADC2        130 71  86  72  72  141 79  66  
DAC1        110 78  105 120 52  120 1500    146 
DAC2        1021    1015    1009    1010    954 1050    1043    1090    
SCALE1      55  53  55  55  56  56  55  57  
OFFSET1     731 825 807 860 790 870 829 797 
BIASCAL     135 1635    1515    90  105 1695    3180    390 
LTEMP       26.50   28.25   28.50   27.00   25.00   26.25   28.00   27.25   
RTEMP       25.25   28.25   29.50   28.50   26.25   26.50   30.00   27.25   
============================================

hartytp commented 6 years ago

@wizath what is PN8V0?

Why P when it's a current? Why not I? :)
IIRC, N8V0 is the bias voltage for the FET, so I can't believe it has >0.5A draw. Am I correct in thinking that it's a typo and that PN8V0 is really I28V0?
If that's correct, why is there so much variance in the bias current? Didn't you tune the bias voltage to give the same bias current on each amplifier channel?

hartytp commented 6 years ago

Power cycled this and looked again:

PGOOD: 1
FAN SPEED: 0 %
AVG TEMP: 30.50 MAX:31.00
CHANNELS INFO
==============================================================================
        #0  #1  #2  #3  #4  #5  #6  #7
DETECTED    1   1   1   1   1   1   1   1   
TXPWR [V]   0.04    0.12    0.05    0.05    0.06    0.05    0.07    0.05    
RFLPWR [V]  0.08    0.04    0.05    0.04    0.05    0.09    0.05    0.04    
TXPWR [dB]  -12.02  -11.85  -13.18  -14.07  -12.36  -14.00  -12.91  -12.51  
RFLPWR [dB] -14.77  -11.63  -15.08  -15.02  -11.54  -14.66  -13.23  -13.64  
P30V [A]    0.111   0.158   0.162   0.142   0.161   0.126   0.159   0.164   
P6V0 [A]    0.022   0.032   0.032   0.028   0.033   0.025   0.032   0.033   
PN8V0 [mA]  0.473   0.673   0.694   0.605   0.701   0.538   0.678   0.696   
ON      1   1   1   1   1   1   1   1   
SON     1   1   1   1   1   1   1   1   
IINT        0   0   0   0   0   0   0   0   
OINT        0   0   0   0   0   0   0   0   
SINT        0   0   0   0   0   0   0   0   
SINTE       0   0   0   0   0   0   0   0   
SINTV       0   0   0   0   0   0   0   0   
OVC     0   0   0   0   0   0   0   0   
ADC1        70  197 82  86  98  86  119 84  
ADC2        131 71  86  73  76  147 80  66  
DAC1        110 78  105 120 52  120 1500    146 
DAC2        1021    1015    1009    1010    954 1050    1043    1090    
SCALE1      55  53  55  55  56  56  55  57  
OFFSET1     731 825 807 860 790 870 829 797 
BIASCAL     135 1635    1515    90  105 1695    3180    390 
LTEMP       26.00   29.25   30.00   29.25   27.00   28.50   31.50   29.25   
RTEMP       27.50   29.50   31.00   29.00   27.50   27.00   30.00   29.25   
==============================================================================

So the currents look quite different

Edit: but after a few minutes of warmup they are returning to their previous values

hartytp commented 6 years ago

The fluctuations in the Pn8V0 reading are surprisingly large. e.g. Cht goes between 0.582 ish and 0.62, which seems like a lot

hartytp commented 6 years ago

Also, even when the device is in standby mode, the current draw on tha trail is 0.4A for one channel

hartytp commented 6 years ago

Other than the above observations, I don't see anything obviously wrong with the diagnostics. @gkasprow @wizath is there anything in particular I should be looking out for? If not, I'll put the VNA back in and see if I can reproduce my previous observation of gain instability

hartytp commented 6 years ago

I'll see if I can reproduce this. In the mean time, @wizath please could you answer my questions above e.g.

what is PN8V0?
how large do you expect the fluctuations on these readings to be?
why is there such a larger difference between the Pn8V0 reading from channel to channel?

wizath commented 6 years ago

what is PN8V0?

It's the current measurement on the negative voltage LDO.

how large do you expect the fluctuations on these readings to be?

Since it's not calibrated, the accuracy is based on 12-bit ADC and 1% resistor.

why is there such a larger difference between the Pn8V0 reading from channel to channel?

That could be because they're calibrated to draw about the same amount of current each and negative voltage that is set varies highly between channels.

hartytp commented 6 years ago

@wizath Can you double check that woth Greg, please. I thought the NV8 rail was only for biasing the FET gate so the current should be close to zero. Are you sure it's not the 28V0 current?

gkasprow commented 6 years ago

There is resistive divider of several tens of Ohms, so the current consumption is several mA.

hartytp commented 6 years ago

@wizath apologies: i'd missed the 'm' there and thought it was in amps.

~In that case, how do I read the 28V bias current? That's most important for me in terms of checking for start up issues.~ oops, that's the P30V row.

hartytp commented 6 years ago

@gkasprow how did you calibrate the FET gate voltage? IIRC you were going to calibrate it to give a fixed bias current, is that what you ended up doing? If so, what was the target current? Why is the current on channel 1 so much lower than the other channels?

hartytp commented 6 years ago

One other question:

when the device is active, I see the following current draws

PGOOD: 1
FAN SPEED: 0 %
AVG TEMP: 30.50 MAX:32.25
CHANNELS INFO
==============================================================================
        #0  #1  #2  #3  #4  #5  #6  #7
DETECTED    1   1   1   1   1   1   1   1   
TXPWR [V]   0.92    0.12    0.05    0.05    0.06    0.05    0.07    0.05    
RFLPWR [V]  0.62    0.04    0.05    0.04    0.05    0.10    0.05    0.04    
TXPWR [dB]  14.00   -11.89  -13.18  -14.07  -12.32  -14.02  -12.87  -12.54  
RFLPWR [dB] 12.75   -11.67  -15.08  -15.02  -11.50  -14.68  -13.19  -13.68  
P30V [A]    0.133   0.157   0.161   0.139   0.140   0.114   0.157   0.162   
P6V0 [A]    0.028   0.031   0.032   0.028   0.028   0.023   0.031   0.032   
PN8V0 [mA]  0.600   0.668   0.683   0.595   0.600   0.483   0.668   0.688   
ON      1   1   1   1   1   1   1   1   
SON     1   1   1   1   1   1   1   1   
IINT        0   0   0   0   0   0   0   0   
OINT        0   0   0   0   0   0   0   0   
SINT        0   0   0   0   0   0   0   0   
SINTE       0   0   0   0   0   0   0   0   
SINTV       0   0   0   0   0   0   0   0   
OVC     0   0   0   0   0   0   0   0   
ADC1        1501    195 82  86  100 85  121 82  
ADC2        1011    72  86  73  80  159 83  65  
DAC1        110 78  105 120 52  120 1500    146 
DAC2        1021    1015    1009    1010    954 1050    1043    1090    
SCALE1      55  53  55  55  56  56  55  57  
OFFSET1     731 825 807 860 790 870 829 797 
BIASCAL     135 1635    1515    90  105 1695    3180    390 
LTEMP       29.00   28.00   29.25   29.25   28.00   31.50   31.00   31.00   
RTEMP       31.00   30.50   30.25   28.00   28.25   29.00   32.25   30.00   
==============================================================================

when its in standby I see

================================================
␛[2J
PGOOD: 1
FAN SPEED: 0 %
AVG TEMP: 31.35 MAX:29.25
CHANNELS INFO
==============================================================================
        #0  #1  #2  #3  #4  #5  #6  #7
DETECTED    1   1   1   1   1   1   1   1   
TXPWR [V]   0.04    0.11    0.04    0.05    0.11    0.05    0.07    0.05    
RFLPWR [V]  0.08    0.04    0.05    0.04    0.04    0.09    0.04    0.04    
TXPWR [dB]  -12.16  -12.04  -13.35  -14.25  -10.77  -14.20  -13.04  -12.63  
RFLPWR [dB] -14.92  -11.81  -15.25  -15.22  -9.89   -14.87  -13.37  -13.77  
P30V [A]    0.022   0.078   0.082   0.074   0.053   0.030   0.065   0.091   
P6V0 [A]    0.004   0.016   0.016   0.015   0.011   0.006   0.013   0.018   
PN8V0 [mA]  0.094   0.330   0.351   0.314   0.221   0.127   0.278   0.387   
ON      0   0   0   0   0   0   0   0   
SON     0   0   0   0   0   0   0   0   
IINT        0   0   0   0   0   1   0   0   
OINT        0   0   0   0   0   0   0   0   
SINT        0   0   0   0   0   0   0   0   
SINTE       0   0   0   0   0   0   0   0   
SINTV       0   0   0   0   0   0   0   0   
OVC     0   0   0   0   0   0   0   0   
ADC1        62  187 73  76  187 75  112 77  
ADC2        124 63  80  64  69  150 72  60  
DAC1        110 78  105 120 52  120 1500    146 
DAC2        1021    1015    1009    1010    954 1050    1043    1090    
SCALE1      55  53  55  55  56  56  55  57  
OFFSET1     731 825 807 860 790 870 829 797 
BIASCAL     135 1635    1515    90  105 1695    3180    390 
LTEMP       30.25   28.00   28.00   29.50   28.00   31.25   31.50   31.00   
RTEMP       28.00   29.25   28.00   27.00   28.25   29.00   29.00   29.00   
==============================================================================

Why is the 30V current draw so high in standby for some channels? It seems like there is only a factor of 2 between normal operation and standby mode. Is this just ADC offsets? If so, how accurate is this measurement?

ghost commented 6 years ago

AVG TEMP: 31.35 MAX:29.25

This looks fishy.

hartytp commented 6 years ago

Oh dear, have we lost David?

hartytp commented 6 years ago

https://github.com/sinara-hw/Booster/issues/36#issuecomment-409612201

@wizath

hartytp commented 6 years ago

@gkasprow @wizath I haven't been able to reproduce this, so maybe it was a measurement error after all. I'm surprised if that is the case, since I took the data quite carefully, and had a colleague present when I observed this issue. Anyway, I'll close this for now. If it is a real issue, I'm sure we'll see it again soon.

jordens commented 6 years ago

@dtcallcock is also metastable now

hartytp commented 6 years ago

which, I'd suggest, is an improvement from his days here

sinara-hw / Booster

Booster: metastability #36