Well,
all are invited to download my evaluation version. It's a full function release with time limitation usage.
There are already various IM freq. build in and additional you may define your own IM cluster using the multi sine generator.
HpW
HpW I looked at your website, and the information is overwhelming. I can believe that for many people it is difficult to know what is what. Lots of pics with no narrative.
I would think that you would get more takers if you had some kind of step by step user guide to explain what is what and in general more user-targeted info.....
Jan
I would think that you would get more takers if you had some kind of step by step user guide to explain what is what and in general more user-targeted info.....
Jan
Jan,
Install the SW and then open the help file where you find the tutorial(s) and much more..
I hope helping you and others too
HpW
Ok, so what sort of generator distortion specs can be expected via the SW/soundcard route?
Also as a measurement package, how does ur SW/soundcard combo compare with the QA?
What have you found to be the best performing soundcard?
Perhaps you might wish to post some results that mirror what has been done so far withe the QA? Unless you also own a QA and can do side by side comparisons?
Also as a measurement package, how does ur SW/soundcard combo compare with the QA?
What have you found to be the best performing soundcard?
Perhaps you might wish to post some results that mirror what has been done so far withe the QA? Unless you also own a QA and can do side by side comparisons?
HpW-
I think software is fine. The hardware issues are the limitation. The residuals I'm looking at with the simple oscillators in the box are at the .003% level and mostly noise. Using external analog generators I got the residual down to -126 and I think that can be lowered significantly. Software brings its hardware limitations. The DAC's I have looked at are no better in this than the existing source but have that pesky computer to deal with. Some external hardware may be able to enhance the PC's performance.
I think software is fine. The hardware issues are the limitation. The residuals I'm looking at with the simple oscillators in the box are at the .003% level and mostly noise. Using external analog generators I got the residual down to -126 and I think that can be lowered significantly. Software brings its hardware limitations. The DAC's I have looked at are no better in this than the existing source but have that pesky computer to deal with. Some external hardware may be able to enhance the PC's performance.
I wanted to share my experience with Quantasylum's customer service.
While testing a power amplifier I must have exceeded the QA400's input voltage and damaged the left channel (my fault). I contacted Matt at Quantasylum asking for advice on repairing it myself. He asked me to send it back to them. They did the repair without cost to me. Turn around time was one week.
I could not be happier.
While testing a power amplifier I must have exceeded the QA400's input voltage and damaged the left channel (my fault). I contacted Matt at Quantasylum asking for advice on repairing it myself. He asked me to send it back to them. They did the repair without cost to me. Turn around time was one week.
I could not be happier.
Regardless of QA's motives it sounds like great customer support.
I have seen equipment where an attenuator constructed from very small wattage resistors was placed before the input, or after the output, to serve as a protective fuse in the event of excessive levels, or accidental connection to someplace where it shouldn't have been connected. Sometimes these attenuators were built into the instrument design, and other times they were a shop-built afterthought in a Pomona box. And, sometimes they were effective, while at other times they weren't.
FM 2-way radio techs who keyed a transmitter into a Cushman signal generator usually didn't do it very often. Don't ask how I know that.
Dale
I have seen equipment where an attenuator constructed from very small wattage resistors was placed before the input, or after the output, to serve as a protective fuse in the event of excessive levels, or accidental connection to someplace where it shouldn't have been connected. Sometimes these attenuators were built into the instrument design, and other times they were a shop-built afterthought in a Pomona box. And, sometimes they were effective, while at other times they weren't.
FM 2-way radio techs who keyed a transmitter into a Cushman signal generator usually didn't do it very often. Don't ask how I know that.
Dale
I have an HP time interval probe that uses a quad mos gate in the probe tip as the sensor. It turns out its easy to change (socketed) which is great because its an almost unprotected mosfet gate input. It was designed that way. I have confirmed that they are fragile and replaceable. . . .
That was my model for using through hole parts on the interface board.
It doesn't look like I'll have much chance to do anything more on my next version of the interface until after CES.
That was my model for using through hole parts on the interface board.
It doesn't look like I'll have much chance to do anything more on my next version of the interface until after CES.
It really depends on the used internal data types: Integer (16/32 bit) or float (64/80bit). Specially when dithering is applied to get ride of quantization errors. I used the highest possible values..
I do not provide any HW (sound card or digital I/O cards), my solution is pure SW based. My plans where to build such HW but the SW requires more attention then I expected
My design/idea was to have an analog asym/sym input/output with attenuation for the input and gain for the output (controlled by a USB connection) and connected to a digital I/O sound card.
This changes from time to time. In the oldest days as I started, there was the ISA Turtle Beach Tahiti (only 16 bit but went down to DC)... currently I use the RME BabyBace (USB with good analog symmetric spec's) and the Lynx Two L22 (PCI based). In case of low freq. or noise measurement special attention is required while the power supply is bellow 100hz usual very noisy.
The next generation sound cards will be using USB 3, but the overall performance will never be better then the used ADC/DAC. The HW manufacturers count here often each cents and this will be the ...
I hope answering your questions ...
HpW
Yes. I built Pete Millett's "Test & Measurement interface for Soundcards"
I could have crossed one of the interconnects. I am not sure exactly when or how it happened. It did not wipe the left channel out completely. It was just showing a severely attenuated (60dB) measurement.
I just have to be more careful in the future. One of the reasons I am following this post is the need for a custom interface for the QA400. I am looking forward to Demian's final version.
Ouch!
The failure was *with* the PMillet interface board inline??
Good case here for color coding your cables - fewer accidental errors.
It's an old trick that I started to use myself in the 70s when wiring insanely complicated rock band stages (nightly) where all the cables had the same connectors and were basically black cable and silver plugs/jacks. Not sure, but I may have been the first to do this on-the-road.
_-_-
The failure was *with* the PMillet interface board inline??
Good case here for color coding your cables - fewer accidental errors.
It's an old trick that I started to use myself in the 70s when wiring insanely complicated rock band stages (nightly) where all the cables had the same connectors and were basically black cable and silver plugs/jacks. Not sure, but I may have been the first to do this on-the-road.
_-_-
The above quote is from post #1. I think Demain has this covered in his design.
I too have a bin full of adapters and just as many different kinds of cables. When testing the workbench can get cluttered with cables on top of and across everything else. Nothing is full proof but I think the 3.5mm connectors will reduce the chance of accidental connections. Sounds like a good idea to me and I plan to use this when I build Demian's interface.
Yes, the Lynx uses the AKM AK5394 ADC. The RME do not use the AK5394, the noise will rise after 48Khz using 192Khz SR.
Keep in mind using Vista / Win 7 and later using the ASIO or WASPAI exclusive mode otherwise MS Window DSP mixers gives the measurement performance the rest
BTW: It would be a great bargain, if someone would start a high performance DIY ADC/DAC project, just using the demo boards as the AKM AK5394, made you nuts do the poor input pcb layout and using the old 5532 op-amps
.
Hp
The AKD5394A demo board can be really good but not necessarily suitable for general purpose performance testing. I replaced the opamps with LME49710's with a substantial reduction in distortion. The clock on it is bad and you need a good external master clock. The big OXCO I'm using now is way overkill (I think). I'll try some tricks with a TentLabs or Crystek when I can get back to it.
I would really like an external module(s) with AK5394A and something like the AK4399 or ESS as a DAC with a more flexible input/output. SPDIF out would really work better for this than USB. No grounding stuff to address. However I don't need it so bad that I'm laying it out right now.
I would really like an external module(s) with AK5394A and something like the AK4399 or ESS as a DAC with a more flexible input/output. SPDIF out would really work better for this than USB. No grounding stuff to address. However I don't need it so bad that I'm laying it out right now.
How's it going -- interface? I know with the holidaze and now CES convention soon, things slow down on personal projects.
My problem is that very few if any "24" bit machines are good to anywhere near the full 144dB dynamic range. More like 18-20 bits.
The notch/null of fundamental helps a lot and so that is a 'must have' to get reasonably accurate data below -90-100dB; If that is even needed by users. And, not only from this device but all of them (ADC) seem to have accuracy issues when trying to use the full 24 bits.
Thx-RNMarsh
My problem is that very few if any "24" bit machines are good to anywhere near the full 144dB dynamic range. More like 18-20 bits.
The notch/null of fundamental helps a lot and so that is a 'must have' to get reasonably accurate data below -90-100dB; If that is even needed by users. And, not only from this device but all of them (ADC) seem to have accuracy issues when trying to use the full 24 bits.
Thx-RNMarsh
Last edited:
Hi Rick,
What we need to know is the effect of noise and input amplifier offset on the ADC input. It's not fair to blame the ADC for inaccuracies unless these effects are well understood. Can you set up a test for this. Would a servo to drive the offset out be of any advantage?
What we need to know is the effect of noise and input amplifier offset on the ADC input. It's not fair to blame the ADC for inaccuracies unless these effects are well understood. Can you set up a test for this. Would a servo to drive the offset out be of any advantage?
Last edited: