Generally it is best to connect the USB directly, without a hub. But there is sometimes a hub internally in the PC.
Short and good quality USB cables are of course the best. I have not seen issues with relatively long cables though, e.g. 1.8 m.
Short and good quality USB cables are of course the best. I have not seen issues with relatively long cables though, e.g. 1.8 m.
Hi Jens,
I'm running mine with a USB cable longer than 1m (probably 2m), it works fine. I did use a really good cable. Windows goes stupid sometimes and I have to reboot the computer, cycle power on the RTX unit and sometimes unplug and insert the USB. This is purely a Windows problem (10 Pro, 7 Pro was solid). I'm using both Arta for quick stuff, and M.I.
You designed a wonderful unit Jens, thank you so much!!
I'm running mine with a USB cable longer than 1m (probably 2m), it works fine. I did use a really good cable. Windows goes stupid sometimes and I have to reboot the computer, cycle power on the RTX unit and sometimes unplug and insert the USB. This is purely a Windows problem (10 Pro, 7 Pro was solid). I'm using both Arta for quick stuff, and M.I.
You designed a wonderful unit Jens, thank you so much!!
What is the input impedance of the RTX analyser input? I am sure this information is out there somewhere, but I don't know where to look.
I am asking because I need to look at a high-voltage output from a an electrostatic headphone amp, which can easily output AC signals at 1 kV (peak-to-peak). The idea would be to make a voltage divider using a resistor and the the RTX input impedance. For example, if the RTX input impedance was 100 kOhm (purely resistive), I could use a 10 MOhm resistor to make 1:100 voltage divider. This would attenuate the amplifier output from 1 kV to 10 V, which is easily in the range that can be analysed with the RTX.
I am asking because I need to look at a high-voltage output from a an electrostatic headphone amp, which can easily output AC signals at 1 kV (peak-to-peak). The idea would be to make a voltage divider using a resistor and the the RTX input impedance. For example, if the RTX input impedance was 100 kOhm (purely resistive), I could use a 10 MOhm resistor to make 1:100 voltage divider. This would attenuate the amplifier output from 1 kV to 10 V, which is easily in the range that can be analysed with the RTX.
I did something similar for my ESLs and the AP analyser. Make sure you also include the input capacitance and use a capacitor in parallel to the large input resistor to get a flat attenuation curve over frequency.
You probably have to tune that parallel capacitor while measuring the frequency response with an audio oscillator.
Jan
You probably have to tune that parallel capacitor while measuring the frequency response with an audio oscillator.
Jan
Perhaps use a 100:1 scope probe that is rated for at least 350Vac out to the frequency you are interested in. The probe would then be designed for flat frequency response out to its bandwidth when used with a nominal 1M loading (with some parallel cap), and have some ability to flatten the frequency response. I do that with a EMU0404 with nominal 1M input, and use the probe compensation and a loopback to get the 'best' response, and then use active REW compensation for a flat response.
I think I would try a 100:1 scope probe, characterise it with a known lower voltage signal. However, I'm not comfortable with depending on the input impedance of the unit. I haven't checked, but it may change with range and you want to work with knowns.
Jan is also right, you have to compensate for stray capacitance in the input network. Your allowable AC voltage will fall with frequency as will your impedance at the probe tip. I would use resistors rated for 3,500 V, you may need to use a capacitor rated high voltage ahead of your variable capacitor to split what your compensation capacitor sees, it will need to be rated for high voltage also.
However you do this, you'll have to characterise your probe before you can trust it for anything.
Jan is also right, you have to compensate for stray capacitance in the input network. Your allowable AC voltage will fall with frequency as will your impedance at the probe tip. I would use resistors rated for 3,500 V, you may need to use a capacitor rated high voltage ahead of your variable capacitor to split what your compensation capacitor sees, it will need to be rated for high voltage also.
However you do this, you'll have to characterise your probe before you can trust it for anything.
Thanks Pierre,
That's one document I didn't download separately.
Keep in mind those are nominal specs. There is some tolerance. The inputs are normally driven by a much lower impedance so they wouldn't affect most equipment. Using a probe to step down is the opposite situation.
That's one document I didn't download separately.
Keep in mind those are nominal specs. There is some tolerance. The inputs are normally driven by a much lower impedance so they wouldn't affect most equipment. Using a probe to step down is the opposite situation.
Thanks @PierreQuiRoule for the spec sheet and everyone for their thoughs. I guess I'll try my scope probe. Hopefully that does not distort too much.
Your probe won't distort, make sure it's x100 and high voltage rated. Your biggest issues will be frequency response and absolute value.
You can also find the information in the manual, which can be downloaded here: https://www.rtx.dk/support/download-center/manuals-and-guides/manuals/test-equipment-manuals/
Nominal values are 100 kohm in parallel with 37 pF.
The 100 kohm will be fairly accurate, since the input attenuator is made with 0.1% resistors. The capacitance may vary a bit more.
You could also make a 10:1 attenuation with 900 k resistors in parallel with around 4 pF. Of course you need high voltage components and NP0 or similar quality capacitors.
Nominal values are 100 kohm in parallel with 37 pF.
The 100 kohm will be fairly accurate, since the input attenuator is made with 0.1% resistors. The capacitance may vary a bit more.
You could also make a 10:1 attenuation with 900 k resistors in parallel with around 4 pF. Of course you need high voltage components and NP0 or similar quality capacitors.
9 x 100K = 900K
9 x 300V = 2.7KV
String them up and put them in a plastic paper binder
Or better still, order more of those Chinese HV resistors.
9 x 300V = 2.7KV
String them up and put them in a plastic paper binder
Or better still, order more of those Chinese HV resistors.
... or even better use a Dale or BC 3500V rated resistor. They come in 1% metal film types. Less stuff strung together, and these have very good voltage coefficients. You can string resistors (and caps) in series to improve the voltage coefficient if it is a problem. I think 9 resistors is a bit much.
A scope probe set up for 1 Meg will be off by 10X at DC with 100K. This will take a few minutes to read and its old but you will understand probes and probing much better: https://w140.com/tekwiki/images/1/19/062-1120-00.pdf This is newer but more superficial (and sales oriented) : https://www.testequity.com/UserFiles/documents/pdfs/tektronix/probe_fundamentals.pdf
What I would recomend for this (and many other tasks) is a probe buffer amp. A dual opamp, CMOS set to 1 Meg + 20 pF and unity gain to drive the audio analyzer. Diodes to limit the peak voltage in. Socketed so easy to replace when zapped. I would use the OPA1656 because its more linear with a high impedance source that most opamps. This PCB would work https://www.ebay.com/itm/3118527310...bbTdhmXi8W9RjDmCdUIctn/xA=|tkp:Bk9SR4zTofSBZA
What I would recomend for this (and many other tasks) is a probe buffer amp. A dual opamp, CMOS set to 1 Meg + 20 pF and unity gain to drive the audio analyzer. Diodes to limit the peak voltage in. Socketed so easy to replace when zapped. I would use the OPA1656 because its more linear with a high impedance source that most opamps. This PCB would work https://www.ebay.com/itm/3118527310...bbTdhmXi8W9RjDmCdUIctn/xA=|tkp:Bk9SR4zTofSBZA
That may seem like a silly question, but... is RTX DAC non-oversampling (NOS) or does it use some sort of aliasing?
What is the most straightforward way on the PC-side for the RTX 6001? Windows or Linux? Arta or REW? Something else?