Well. I know that 99% of the people around prefer to talk about HW, LDOs, silver micas, cables, asf... 😉
How about a little SW talk? 😉
How about a little SW talk? 😉
Had a quick look at the python library. Resisting the urge to vomit at the sight of Python..
That library simply operates by automating the umaanger serial terminal commands to the dam umanager. If the commands in the manuals are identical then it will work without a problem - there may be some parameter value differences though.
The change you need to be careful of the version of firmware that you would use for uploading to the dam itself (the "download" command). The Spartan6 fgpa firmware would work differently (both in the sign magnitude) but also operating the pinouts for the 4 R2R ladders - using the wrong version of the firmware could cause damage to the dam.
So use the wrong firmware uploaded means you could cause permanent damage.
That library simply operates by automating the umaanger serial terminal commands to the dam umanager. If the commands in the manuals are identical then it will work without a problem - there may be some parameter value differences though.
The change you need to be careful of the version of firmware that you would use for uploading to the dam itself (the "download" command). The Spartan6 fgpa firmware would work differently (both in the sign magnitude) but also operating the pinouts for the 4 R2R ladders - using the wrong version of the firmware could cause damage to the dam.
So use the wrong firmware uploaded means you could cause permanent damage.
Just looking:
It looks like range(4) (ie 1-4) for the source validation may prevent 5 "i2s" and 7 "auto" being used.
There's also no crossfeed and phase mode support for the new feature - this would be relatively simple to add new support.
Code:
VOLUME_INF=-80
VOLUME_SUP=10
VOLUME_POT=-99
INPUT_SRC_SET=range(4)
FSET_EXT_STR_TO_INT_D = dict(linear=4,mixed=5,minimum=6,soft=7)
FSET_EXT_NUM_TO_INT_D = {'1':4,'2':5,'3':6,'4':7 }It looks like range(4) (ie 1-4) for the source validation may prevent 5 "i2s" and 7 "auto" being used.
There's also no crossfeed and phase mode support for the new feature - this would be relatively simple to add new support.
The DAM has minimal format/syntax check so be sure you do what you should or things may go wrong. This is experience talking. But hey, you use a computer - what could go wrong 😉
//
//
When I think of all of the things I did to my first SOEKRIS board I am amazed that I was able to mess it up with the removal of the ceramic capacitors. Everything now is about half the size; the first board was much easier to play around with due to the larger components.
I did not use heat - sharp dykes in the middle of the cap and then carefully squeeze. Seemed to go uneventfully … I cleaned two of the pads (left channel) and they cleaned up easily but it did not seem worth the trouble. The pads are there and undamaged.
The left channel plays music but there is something very wrong.
Everything looks fine on he board.
The right channel seems to be fine. Did not listen to it long so I have no idea what the removal did for the sound.
There is this noise, in the 120 hz region, not LOUD - you can hear the music on top of it easily but no one would want to listen to it like this.
Any ideas or is this now a mono DAC?
Not that I was not warned not to take things off of the board by the manufacturer. Wish I had left it alone. Never thought of myself as a cat but I must have something in common with them. Luckily, I just killed the board.
Yet I know if I had it to do over again I would have done it, again. Got worried Peter and analog_sa would think there was some veiled "blaming" them. They certainly did not recommend clipping them out and I suspect, too late, that, like most shortcuts, has its pitfalls. Get out the soldering irons!
I did not use heat - sharp dykes in the middle of the cap and then carefully squeeze. Seemed to go uneventfully … I cleaned two of the pads (left channel) and they cleaned up easily but it did not seem worth the trouble. The pads are there and undamaged.
The left channel plays music but there is something very wrong.
Everything looks fine on he board.
The right channel seems to be fine. Did not listen to it long so I have no idea what the removal did for the sound.
There is this noise, in the 120 hz region, not LOUD - you can hear the music on top of it easily but no one would want to listen to it like this.
Any ideas or is this now a mono DAC?
Not that I was not warned not to take things off of the board by the manufacturer. Wish I had left it alone. Never thought of myself as a cat but I must have something in common with them. Luckily, I just killed the board.
Yet I know if I had it to do over again I would have done it, again. Got worried Peter and analog_sa would think there was some veiled "blaming" them. They certainly did not recommend clipping them out and I suspect, too late, that, like most shortcuts, has its pitfalls. Get out the soldering irons!
Sounds like only signal tracing with an oscilloscope is going to be able to answer that. 🙁
At least somebody did that library job (FOC). 😉
A job that could (should?) have been done by the manufacturer. 😛
There are $50 RPi DAC HATs out there that come with full blown drivers.
Volume control, mute, changing filters, changing outputs all these basic operational tasks I'd prefer to run via USB.
Anyhow.
Thx for pointing into potentially critical areas. I hope that the DAC is well documented. That should make it rather easy to get the library adapted.
And for me most important to know: As is, that python toolbox is not compatible.
And that could put quite some extra load on my shoulders. 🙄
A pity, as you mention, that the DAC allows to load firmware that might harm the device. That I'd consider a really weak spot.
Thx.
Just learning about bypass and ringing - this is the same as the oscillations for AC snubbing I assume, but given the square wave on the levels at 3MHz does the output have ringing and need a number of values of bypass capacitors - how is this tacked on the switches as I think I can only see one cap.
Does the slew rate allow a lower rate?
Does the slew rate allow a lower rate?
The "ringing" talked about re: DAC is not the same as the one a snubber is curing. DAC ringing is (mostly) due to it's digital filter and is a consequence of software calculations while a snubber cures an oscillation in a group of electrical components. The cap on the output is not a bypasss cap (i.e. a cap in series) but a decoupling cap, or a low pass filter if you like. This cap position could indeed remove "ringings" if it was set in low enough. But it is there to filter out high frequency energy way above the audible range.
Your second question does not make sense.
It seems that you are not skilled enough yet to phrase the correct questions. I kindly suggest that you read up on your used concepts on e.g, wiki etc.
//
Your second question does not make sense.
It seems that you are not skilled enough yet to phrase the correct questions. I kindly suggest that you read up on your used concepts on e.g, wiki etc.
//
Last edited:
So the power going to the LVC switches is a digital pulse at a frequency of 3MHz so the leading edges of the switching has ringing. Only a single is required? Does the speed of the LVC slew to switch reduce ringing?
I'm happy with the digital software based ringing from approximating a square wave with a limited set of frequencies, and FIR/IIR for realtime image processing. I never shoot my staff for asking a question.
OK - maybe I misinterpreted your questions - sorry. I don't think I can help you. I'm sure other have ideas. And, it wasn't a shot.
//
//
In earlier posts (re 1021) there is a lot of mention of improving ''vref'', im not quite clear if this +/-4V supply is the same vref on dam1121/1941?
I assume it has to be because there is not much else in the analogue supply area aside from the op amp regs.
In that case I seen soekris mention the tolerance between positive and negative vref from the regs, small imbalances caused increases in distortion.
How would you keep +/- voltages closely matched and also sustained with battery supply?
Last edited:
An excellent point and one i believe has been raised before, perhaps even by Søren. Only, not having a personal interest in batteries i have forgotten all about it. Indeed, without a common reference this cannot possibly work well.
Thank you for correcting me.
Thank you for correcting me.
Imbalance in vref+ and vref- will create more 2nd harmonic distortion, which some people might like 🙂 Batteries with same age and charge level will typically be very close....
Agreed, seems odd when you look at the varying power levels of batteries, the non-linear discharge vs time.
The switches (https://assets.nexperia.com/documents/data-sheet/74LVC595A.pdf) they do state variations in transition times based on voltage.. however I strongly doubt that someone could hear that and if they do it's more likely to some other aberration or absence of aberration such as noise from a PSU.
Also adding caps in the supply only provides filtering and ripple-countering pool - with the speed of switching over 3Mhz you're starting to need to think about high frequency circuit design (bypass caps to protect against ringing etc). Although I suspect there's more issues that would cause more noticeable aberrations.
Sorry but I'm a numbers guy - so if it shows on the oscilloscope then there's a difference. I want the dam to give a precise and accurate representation, I can then choose components to add aberrations such as tube amps etc.