Current Server Time: 11:59 (Central Europe)

#apertus IRC Channel Logs

2019/06/23

Timezone: UTC


01:06
illwieckz
left the channel
01:07
RexOrCine
changed nick to: RexOrCine|away
05:34
Bertl_oO
off to bed now ... have a good one everyone!
05:34
Bertl_oO
changed nick to: Bertl_zZ
08:23
BAndiT1983|away
changed nick to: BAndiT1983
08:41
se6astian|away
changed nick to: se6astian
08:48
se6astian
good day
10:39
se6astian
changed nick to: se6astian|away
10:40
se6astian|away
changed nick to: se6astian
11:03
Y_G
joined the channel
11:20
Y_G
Hi se6astian, I had some questions about the setting the analog gain (PGA).
11:20
Y_G
In the `set_gain.sh`, along with setting the analog gain(reg 115), the register for ADC_range(116), Adc_range_mult2(reg 100) and offset(reg 87 and 88) are also being set.
11:20
Y_G
In daemon implementation do we want something like `set analog_gain` to do all these things or just set the register for analog gain (115).
11:35
BAndiT1983
changed nick to: BAndiT1983|away
11:35
BAndiT1983|away
changed nick to: BAndiT1983
11:52
BAndiT1983
hi Y_G, if i remember correctly then the script was meant like a preset first and values selected fitting it
11:52
se6astian
`set analog_gain` to do all these things <- that
11:54
BAndiT1983
// Set division first, to prevent overwriting gain values
11:54
BAndiT1983
SetConfigRegister(115, _divPGA[_analogGainValue]);
11:54
BAndiT1983
SetConfigRegister(115, _gainPGA[_analogGainValue]);
11:54
BAndiT1983
// SetConfigRegister(100, 1);
11:54
BAndiT1983
// SetConfigRegister(87, 2000);
11:54
BAndiT1983
// SetConfigRegister(88, 2000);
11:54
BAndiT1983
this was in previous version of CMV12000Adapter, now CMV12000Module
11:54
BAndiT1983
diff -> https://github.com/apertus-open-source-cinema/axiom-control-daemon/commit/4ea5116cefc38b418e565d118f57422feca599e9#diff-96249d75c14233a0fdad079c414dc755
11:58
BAndiT1983
there are 2 ways to solve it, add the lines again, which were removed from analog_gain or test set_register and re-create the script with it
12:03
aombk
left the channel
12:06
Y_G
I think adding the lines back should be sufficient
12:08
aombk
joined the channel
12:14
BAndiT1983
then the infos i had back then were different, from what it's now
12:14
se6astian
sounds good
12:14
BAndiT1983
so, off for a moment, will be back a bit later
12:15
BAndiT1983
changed nick to: BAndiT1983|away
12:25
Y_G
left the channel
12:47
illwieckz
joined the channel
12:53
Y_G
joined the channel
12:53
Y_G
left the channel
13:08
Bertl_zZ
changed nick to: Bertl
13:08
Bertl
morning folks!
13:08
apurvanandan[m]
Good morning
13:11
apurvanandan[m]
Do the SERDES primitives adjust the clock in source synchronous systems to the middle of the eye?
13:11
apurvanandan[m]
Like the ISERDES
13:18
Bertl
no, you need to do any adjustments with timing constraints or actively with delay elements
13:20
apurvanandan[m]
Ok, great info :)
13:25
Bertl
typically you want to use the same I/O element to generate the clock
13:26
Bertl
so for example, if you use a DDR I/O element for your data, you also use one to generate the clock
13:27
Bertl
if you utilize an OSERDES for the data, you also want to use an identically configured OSERDES for the clock signal
13:27
Bertl
this ensures that all the timing and phase relations are identical
13:27
apurvanandan[m]
ahh, I see
13:30
apurvanandan[m]
For the data, I used two OSERDES, a master with a slave SERDES. So, do I need to use two SERDES again for the clock?
13:32
Bertl
that would be the 'smart' thing to do :)
13:33
apurvanandan[m]
Okay :)
13:47
illwieckz
left the channel
13:51
illwieckz
joined the channel
14:03
aSobhy|away
changed nick to: aSobhy
14:51
Bertl
hey aSobhy! How's it going?
15:07
illwieckz_
joined the channel
15:09
illwieckz
left the channel
15:34
aSobhy
I'm very delayed :/
15:34
aSobhy
I'm stuck-ed at receiver side at the deserializer module,
15:34
aSobhy
I left that part and simulating the rest
15:34
aSobhy
(write a small cpp file to get the seed with longest words that will lead me to a non reset "00000000" before looping again)
15:34
aSobhy
I almost finished simulation and I'll be back to the deserializer and start docummentation
15:36
Bertl
hmm, are you talking about the PRNG here?
15:38
aSobhy
yeah the cpp for PRNG
15:40
Bertl
well, there are well known polynomials with length 2^n-1
15:42
Bertl
https://www.maximintegrated.com/en/app-notes/index.mvp/id/4400
15:42
Bertl
for example lists them
15:43
Bertl
https://en.wikipedia.org/wiki/Linear-feedback_shift_register#Some_polynomials_for_maximal_LFSRs
15:44
Bertl
so for those you always know that they will produce 2^n-1 values before returning and thus reach each and every state except for '0' or '~0' depending on the type of LFSR
15:45
aSobhy
I used 8-bit shift with bits 5.3.2.0 with seed "11100111 and found a 254 word generated before reaching the reset state "00000000"
15:48
aSobhy
254 out of 256 possible words
15:48
Bertl
okay, but why do you want that?
15:49
Bertl
if you look at the list (I posted) you see that the polynom 6.5.4.0 for 8bit will give you 255 values
15:51
aSobhy
It has no purpose that's what I thought
16:02
aSobhy
okay I changed it it gives 255 with no zeros :)
17:22
illwieckz_
left the channel
17:35
illwieckz_
joined the channel
18:19
BAndiT1983|away
changed nick to: BAndiT1983
19:04
niemand
joined the channel
19:07
aombk
left the channel
19:07
aombk
joined the channel
19:29
Y_G
joined the channel
19:30
aombk2
joined the channel
19:31
BAndiT1983
changed nick to: BAndiT1983|away
19:32
aombk
left the channel
20:10
niemand
left the channel
21:09
se6astian
off to bed
21:09
se6astian
good night
21:09
se6astian
changed nick to: se6astian|away
21:20
illwieckz_
changed nick to: illwieckz
21:35
Y_G
left the channel
21:47
Bertl
off for now ... bbl
21:47
Bertl
changed nick to: Bertl_oO