PRP1 - Low cost 2GHz power rail probe

[kw1]

not in stock now from Eleshop.eu
Expected on 20 Mar

eleshop can start delivering today   

Klaus

[Filippo52]

What a shame!!
Last night, after FiveFiveFive announced it, the discount also appeared on Aliexpress Italy, and indeed, the last one I got was €133.50. That's €10 less than Eleshop, and above all, until last night it wasn't in stock on Eleshop, so the delivery date wasn't known.
This morning it appeared in stock. Who knows, maybe I would have gotten it on Eleshop, but €10 more isn't entirely negligible.

Now the situation is reversed: on Aliexpress it's "out of stock," while it seems to be available on Eleshop.

Delivery on Aliexpress is scheduled between March 27th and April 7th; they usually arrive sooner, but during this period I've seen items sitting at airports in China, waiting for over a week.

[rogerggbr]

I would have preferred to buy it from Eleshop as they are just across the waves in the Netherlands, but unfortunately their minimum €170 for UK orders makes it rather expensive to use them if there's nothing else you need to buy. The price from AliExpress is comparable to the PRP1 Eleshop price (once you've done €->£ and added on P&P and VAT), even without the anniversary voucher, so it was an easy win for me, even if the delivery time will be slower.

The other downside of buying from Europe into the UK at the moment is you need to add fees and taxes charged by the carrier, which end up anything from 20 to 30%. Thanks Brexit. Done it three times now in the last few years and it still comes as a shock. This is assuming the carrier is one of the biggies e.g. UPS/Fedex/DHL etc who do it properly. I'm not blaming the shop/seller, it's the system.

[kw1]

Low frequency transfer function:
Note this measures up to 10MHz, which includes the low frequency crossover point of the probe.


Hi tszaboo,
I received my PRP1 this week, and I'm already getting along with it just fine. Great device!

However, I'm still unclear about the following:
Why is the rise time of the rising edge so smoothed out starting from the middle (pict.1-3)?  (9 ns from original to 2.9 µs with PRP1)

The measurement was performed using an Fgen. SDG 2042x (50 ohms, 100mV PP) and an Osc. SDS2104x with p57 Adapter.
The amplitude decreases also with higher frequency (approx -6db) (pict 4)
In your image, the bandwidth looks very linear.

What am I doing wrong?


klaus

[tszaboo]

Hi Klaus. Thanks for buying the PRP1. I made this measurement myself, and it was a bit of head scratcher at first.
But it makes sense. What you see is the change in input impedance over the frequency range.
At low frequency the probe impedance is 50 KOhm, and it decreases to the nominal ~50 Ohm with higher frequency. The probe is also designed to work with ~0Ohm input impedance, and generators have 50 Ohm output. So it's the input signal that changes it's shape. If you want to verify the frequency response, I suggest to use the test setup below, to go to ~1MHz.

Resulting this response:

Above 1MHz, I didn't find this test setup to be accurate enough, ground loops make a larger error than the probe itself.
If you want to see more about the performance verification, it's published, results are in a the other folder:
https://github.com/tszaboo/PRP1/blob/main/Design%20verification/Test%20description.md

[tautech]

...............
Resulting this response:
............

Pro tip, use the Save button on the front panel which will keep menus displayed rather than the Save/Recall menus in the Utility dropdown. 
If using the Webserver to capture screenshots use the virtual button Save Screenshot that will pop the image to your PC's Downloads folder.

[tszaboo]

...............
Resulting this response:
............

Pro tip, use the Save button on the front panel which will keep menus displayed rather than the Save/Recall menus in the Utility dropdown. 
If using the Webserver to capture screenshots use the virtual button Save Screenshot that will pop the image to your PC's Downloads folder.

Hey this is at least not a polaroid shot scotch taped in place into a research paper. 

[kw1]

Hi tszaboo, thanks so much for the quick reply.

I took a very close look at your test setup.
It’s basically the same as mine, just without the T-connector. Your generator also has 50 ohms, just like mine.

What I don’t understand:
If I omit the T-connector in this test setup, I get an amplitude response of -6 dB.
With the T-connector, the response is almost identical to yours.

where does the zero ohms in your measurement setup come from? It can't be the probe!?


klaus

[tautech]

@kw1
You have a custom probe attenuation set on Ch2......
Maybe this accounts for your result ? 

[tszaboo]

Hi tszaboo, thanks so much for the quick reply.

I took a very close look at your test setup.
It’s basically the same as mine, just without the T-connector. Your generator also has 50 ohms, just like mine.

What I don’t understand:
If I omit the T-connector in this test setup, I get an amplitude response of -6 dB.
With the T-connector, the response is almost identical to yours.

where does the zero ohms in your measurement setup come from? It can't be the probe!?


klaus

The bode plot has the following name, if I understand you correctly:
* Pict2-koax 50ohm term-instead of pp215 Porbe.png
The PRP1 should always have 50 Ohm termination on it's output. Either use the 50 Ohm inside your oscilloscope, or an external one.

On it's input the AWG has a 50 Ohm source impedance. This isn't optimal, and that what changes the frequency response.
Lecroy explains it in one of their manuals:

This is normal, and I have documented it:
https://github.com/tszaboo/PRP1/blob/main/User%20manual/User%20manual.md#source-impedance


Zero ohms is the optimal source impedance. A power supply will have this, which this probe is designed to measure. Just connect it straight to a power supply, and you get all the bandwidth.

[kw1]

...yes this is the 1.3:1 correction for the prp1...

klaus

[kw1]

Zero ohms is the optimal source impedance. A power supply will have this, which this probe is designed to measure. Just connect it straight to a power supply, and you get all the bandwidth.

Hi tszaboo, thanks so much for your explanation... I get it now 

I also ran a test using the SAA30xx Spectrum Analyzer, and the results are excellent.

klaus

[Filippo52]

Sorry, but I see there are people here who are very familiar with the technique used by this probe, and perhaps even the person who "invented" it. I'd like to ask a few questions about safety during use. The goal is to avoid damaging the oscilloscope's 50-ohm input, which has a very low limit of 5 Vrms in the case of the Rigol MHO954; or even damaging the probe.

I have a hunch that, since it's powered by a 5-volt USB and has a DC input with a 50 kOhm rating, and considering there's a capacitor in the path from the rail's DC voltage to the 50-ohm input, there's no way to damage the oscilloscope if you connect it with the knobs not preset. The knobs don't have a voltage scale that could allow for safety presetting, and I assumed such a scale wasn't there because it's unnecessary.

However, I saw in the instructions that there are recommendations for avoiding harmful errors.
So I'm a bit confused and don't want to damage my oscilloscope.

I also read that it's recommended to ground the probe, to the same ground as the oscilloscope and DUT. But I don't have a ground at my house; I live in a city in a building built in 1960, and there's no ground system. Can I use the PRP1 probe, or will I be forced to resell it?

I'd appreciate it if anyone can provide me with information. I'll receive the probe in a few days, and I don't want to damage either the probe or my expensive oscilloscope.

[tszaboo]

Sorry, but I see there are people here who are very familiar with the technique used by this probe, and perhaps even the person who "invented" it.

Thanks but I think Tektronix was who made the first of this probe type.
I just designed a probe with the same principle. Thank you for buying it.

I have a hunch that, since it's powered by a 5-volt USB and has a DC input with a 50 kOhm rating, and considering there's a capacitor in the path from the rail's DC voltage to the 50-ohm input, there's no way to damage the oscilloscope if you connect it with the knobs not preset. The knobs don't have a voltage scale that could allow for safety presetting, and I assumed such a scale wasn't there because it's unnecessary.

There is an AC capacitor in the high frequency path of the composite amplifier. The low frequency path has a power supply that is less than 5V, so it cannot output 5V into your scope. Normally you end up with about +/- 400mV max signal level (500mV after 1.3:1 probe compensation) into 50 Ohm.

I also read that it's recommended to ground the probe, to the same ground as the oscilloscope and DUT. But I don't have a ground at my house; I live in a city in a building built in 1960, and there's no ground system. Can I use the PRP1 probe, or will I be forced to resell it?

I'd appreciate it if anyone can provide me with information. I'll receive the probe in a few days, and I don't want to damage either the probe or my expensive oscilloscope.

The warnings are there, so someone doesn't start measuring for example on the primary side of a power supply, it's not safe for that, they should use a differential probe. Or they disconnect the protective earth (of just the oscilloscope) to turn a scope into a differential probe.

In your situation, I recommend using the USB port on the oscilloscope, and the regular SMA-BNC cable. And when you connect the PRP, you will automatically have the grounds connected, even if you don't have protective earth. It's not more dangerous than a regular scope probe.
 

[bozidarms]

Hi tszaboo,

many thanks for the great product 

[tszaboo]

The stock ran out again.
If someone wants to buy one, write me and we arrange something, I think I still have one from the original production run.

[Filippo52]

Good evening,
My PRP1 has cleared customs for import into Europe, so it should be here in a few days. In the meantime, I had some questions. Forgive me if they seem strange or even silly.

When I measure the PRP1, the life of my oscilloscope's 50-ohm input depends on the proper functioning of the SMD capacitor that isolates/filters the DC voltage. If that capacitor goes flat, the oscilloscope will be seriously damaged.
1 - Question: Is it possible to double that capacitor, obviously increasing its capacitance, and connect it in series to double the protection?
2 - To further increase safety, is it possible to add RF Schottky diodes as clamps before the output to the oscilloscope? I know that the pf would limit the bandwidth anyway, but I don't need 2 GHz, and if desired, they could be inserted as needed, not permanently soldered.

Last question, even if this is just out of curiosity. I bought the probe, so the manufacturer can rest assured.

If I primarily need to detect ripple in switching power supplies, can I make a minimal, passive probe that basically just has a decoupling capacitor?
With a 10 or 22 uF 100V capacitors, I'd get good readings starting from a few hundred Hz and up.

Thanks

[tszaboo]

1 - Question: Is it possible to double that capacitor, obviously increasing its capacitance, and connect it in series to double the protection?

The capacitor is rated for 50V. I guess you can double it, but I doubt you can do that with the same flat frequency response.

2 - To further increase safety, is it possible to add RF Schottky diodes as clamps before the output to the oscilloscope? I know that the pf would limit the bandwidth anyway, but I don't need 2 GHz, and if desired, they could be inserted as needed, not permanently soldered.

You can, the output is a 50 Ohm transmission line. So technically you can use any RF limiters or attenuators, if you wish.

[Filippo52]

Good morning,
My PRP1 arrived from Aliexpress.
The package, although homemade, was sturdy, and there was no damage, not even to the outer cardboard box.
Unfortunately, I saw that it's currently sold out on both Aliexpress and Eleshop. Aliexpress sold a total of 8 of them. It's clear they're produced in small numbers. But they sell out quickly, and hopefully they'll make more. New ones are expected in early May.
The probe works perfectly, and tomorrow I'll try to see if, in my home environment, which is unfavorable due to pollution and electrical system conditions, it can show me something more and better than standard probes on the power rail of some switching power supplies I have at home.
For now, thanks to whoever made this possible at a price that's finally reasonable.
Probes, in my opinion, are generally overpriced.

Filippo

[tszaboo]

If someone got a PRP1 probe, could you please post an scope plot of the self noise of the probe itself?
I've received some images, which show the probe to be more noisy than the specification and my measurements.
At this point I'm not sure which value would be correct. I used my best available scope to determine this noise, and the test setup below:

[scottapotamas]

Mine is from your first batch, so not sure if this is what you're needing.

• DP832 CH1 as supply, 5V setpoint, draw is ~70mW.
• My 50R termination is internal to the scope. Channel attenuation is 1x so you need to apply the PRP gain to the stats.
• Couldn't find my SMA short so used a BNC short (202114) with a adapter to SMA (29-3855).
• CH3 has a BNC short connected, in 50R internal mode, same range etc.
• With no offset applied (at detents) it sits at 277mVavg therefore not clipping. Then zeroed it 'close enough'.
• A normal unrestricted 2GHz 8-bit capture and oversampled "12-bit" which the scope states has -3dB @ 125MHz.

If you want specific settings on the scope or different tests/analysis I'm happy to grab them now.

[tszaboo]

Mine is from your first batch, so not sure if this is what you're needing.

Thanks. The First batch and the Zeenko units are produced with the same BOM, and I tested the Zeenko samples that they sent me.
The setup looks OK, but could you please change the VRMS to AC-VRMS or STDEV? With the DC-VRMS the DC value dominates the measurement. Otherwise I cannot really tell the two traces apart, which is good news.
I have some scope plots from the new 12 bit Rigols, and on the 200uV range it seems very noisy, a lot more than on the 1mV range.

[scottapotamas]

The setup looks OK, but could you please change the VRMS to AC-VRMS or STDEV? With the DC-VRMS the DC value dominates the measurement.

Sorry about that.
Attached, with best-effort PRP1 zero this time.

I know from my SCPI documentation pipeline that 5mV and 10mV have identical internal readings, and the scope automatically throws the 25MHz filter on at 2mV for good reason... I suspect I could get it down by reworking the AC supply like people have for some of the other units but I'm rarely at this end of the scale.

[Lexy]

Here are my measurements.
The setup i used:

• Siglent SDS3034X-HD
• Scope Bandwidth: 350Mhz
• Low noise Power Supply (<50uVrms@20MhzBW)
• Both channels have 50Ω switched on
• C1: PRP1 connected via 30cm sma cable+sma=>bnc
• C3: Used as noise floor reference. 50Ω terminator (TEK)
• SMA short to PRP1 input

Here is the measurement:


For reverence, i also used the scope internals USB output as Power Supply:


We do see that the noise with the scope internal USB PS is slightly higher, but not significant.

Hope this helps,
Alex

[tszaboo]

The setup looks OK, but could you please change the VRMS to AC-VRMS or STDEV? With the DC-VRMS the DC value dominates the measurement.

Sorry about that.
Attached, with best-effort PRP1 zero this time.

I know from my SCPI documentation pipeline that 5mV and 10mV have identical internal readings, and the scope automatically throws the 25MHz filter on at 2mV for good reason... I suspect I could get it down by reworking the AC supply like people have for some of the other units but I'm rarely at this end of the scale.

Thanks the noise is practically zero for 8 bit scopes by the looks of it.

Here are my measurements.
The setup i used:

• Siglent SDS3034X-HD
• Scope Bandwidth: 350Mhz
• Low noise Power Supply (<50uVrms@20MhzBW)
• Both channels have 50Ω switched on
• C1: PRP1 connected via 30cm sma cable+sma=>bnc
• C3: Used as noise floor reference. 50Ω terminator (TEK)
• SMA short to PRP1 input

Here is the measurement:
(Attachment Link)

For reverence, i also used the scope internals USB output as Power Supply:
(Attachment Link)

We do see that the noise with the scope internal USB PS is slightly higher, but not significant.

Hope this helps,
Alex

Thank you Alex this helps a lot.
Even if I take the worst measurement of this with the probe (102uV ) and compare it to the best measurement without (80uV) the calculated noise is 63uV (sqrt(102*102-80*80)) which is less than specifications (65uV)
By the looks of it, the Rigol MHO900 series makes the probe a bit more noisy than the specifications though. I wonder why is that?