Unknown PPS Reading using Tektronix 2235 Oscilliscope

Thread Starter

Geographer0912

Joined Nov 19, 2018
7
Hello all,

I am in the process of hooking up a LiDAR to a GPS and have some concerns about my PPS signal. I am not very experienced in electronics and can use some help.

For the LiDAR to work, it requires an NMEA sentence as well as 1PPS with a pulse width of 100ms. Positive/Negative pulse polarity can be adjusted accordingly. These settings are configurable in the software that came with the GPS/LiDAR.

When reading the signal out of the GPS on the PPS output pin using a Tektronix 2235 Oscilliscope, I receive the signal in the attached image "TIMINGIO_PPS_Edit.png". It occurs once per second. I have added some annotations to the image to help describe what is happening as well as the settings that were used.

Does this PPS signal make sense to any of you? What could potentially be causing this shape?

The attachment "ANTARIS_PPS_Edit.png" is a reading of the PPS signal coming from a test GPS, in the correct signal characteristics as needed by the LiDAR. Once again, I have added some annotations. I hope this helps confirm that my Oscope is not broken or otherwise.

We are in the process of obtaining a digital oscilloscope to confirm the signal but I'd still like to hear any suggestions you all may have in case a differenc oscope produces the same results.


Thanks in advance.
 

Attachments

ebp

Joined Feb 8, 2018
2,332
The PPS signal is differentiated - it is coupled through a small series capacitance to a circuit of low resistance (small and low are relative terms here). It's hard to resolve accurately, but my guess is the time constant (tau = RC, where R is in ohms can C is in farads) is around 1 to 2 milliseconds (3 time constants will bring the signal amplitude back to within about 5% the starting baseline). A 1 ms time constant might be done with 100 nanofarads and 10k ohms or 10 nF and 100k - anyone's guess without knowing some more about the circuit.
 

Reloadron

Joined Jan 15, 2015
7,501
For the LiDAR to work, it requires an NMEA sentence as well as 1PPS with a pulse width of 100ms. Positive/Negative pulse polarity can be adjusted accordingly. These settings are configurable in the software that came with the GPS/LiDAR.
1 PPS is one pulse per second and with a pulse width of 100 mSec that is a 1 Hz rate with a 10% duty cycle. On your TEK 2235 if you set your time base to .2 Sec/Div (next to the slowest sweep speed. then 5 divisions will represent 1 Second and your 100 mS pulse width will be 0.5 Div. Set the scope for a Normal Trigger on the + slope. Adjust the trigger level for a good repeatable trigger.

What I see in your image is a 100 mS negative pulse duration, did you want a 100 mS negative pulse? I can't tell the pulse repetition rate from the image.
Positive/Negative pulse polarity can be adjusted accordingly.
Not quite sure what that means either. I understand it but just do not get what they are getting at.

Also the signal on the left shows AC coupling of the scope for a slow 1 PPS you want to DC couple the scope vertical input as mentioned by ebp.


Ron
 

ebp

Joined Feb 8, 2018
2,332
The time constant is way too short for normal AC coupling of the scope. The spec for that scope is 3 dB down at 10 Hz or less so there would be tailing very much longer than what is seen.
 

Reloadron

Joined Jan 15, 2015
7,501
The time constant is way too short for normal AC coupling of the scope. The spec for that scope is 3 dB down at 10 Hz or less so there would be tailing very much longer than what is seen.
I have one of those scopes (Tek 2235) laying here and ran a 1 Hz 10% duty cycle into it and my screen looks just like the picture when I AC couple. So I figured the image on the left the scope was using or set for AC coupling. So while I agree with you it's not what I am seeing.

Ron
 

Thread Starter

Geographer0912

Joined Nov 19, 2018
7
The PPS signal is differentiated - it is coupled through a small series capacitance to a circuit of low resistance (small and low are relative terms here). It's hard to resolve accurately, but my guess is the time constant (tau = RC, where R is in ohms can C is in farads) is around 1 to 2 milliseconds (3 time constants will bring the signal amplitude back to within about 5% the starting baseline). A 1 ms time constant might be done with 100 nanofarads and 10k ohms or 10 nF and 100k - anyone's guess without knowing some more about the circuit.
Thanks ebp. I will need to do some more research to know what I am working with and how all of this plays it's role. Still, should I be expecting to see a wave form similar to the right image for the signal on the left?

As far as the right signal (ANTARIS_PPS_Edit.png), I measured it using 5v/div. Perhaps this is an error with parameters on the oscope that I have not played around with.

1 PPS is one pulse per second and with a pulse width of 100 mSec that is a 1 Hz rate with a 10% duty cycle. On your TEK 2235 if you set your time base to .2 Sec/Div (next to the slowest sweep speed. then 5 divisions will represent 1 Second and your 100 mS pulse width will be 0.5 Div. Set the scope for a Normal Trigger on the + slope. Adjust the trigger level for a good repeatable trigger.

What I see in your image is a 100 mS negative pulse duration, did you want a 100 mS negative pulse? I can't tell the pulse repetition rate from the image.


Not quite sure what that means either. I understand it but just do not get what they are getting at.

Also the signal on the left shows AC coupling of the scope for a slow 1 PPS you want to DC couple the scope vertical input as mentioned by ebp.

Ron
Thanks Ron. The right image I wasn't so concerned about as I could clearly see what was going on with the wave. You are correct though, 1PPS, 100ms pulse width, negative polarity.

I have attached another image to this reply to show you the results. This is the same signal as mentioned in the image" TIMINGIO_PPS_Edit.png" I apologize for the poor quality image (and rough edits). For this setup I changed it to Normal Trigger, DC coupling, 2v/div and 0.2sec/div. The period looks to be 1 second (as expected) with a voltage of +/- 3v.

In terms of what I meant by "Positive/Negative pulse polarity can be adjusted accordingly. ", I mean that I can choose to send a positive/negative pulse from my GPS and also choose on the LiDAR if the incoming PPS is positive or negative.
 

Attachments

ebp

Joined Feb 8, 2018
2,332
Ron:
If his deflection factors are as marked on the image, the time constant can't be more than a couple of milliseconds. Visually, it is back at zero in well under 10 ms. -3 dB at 10 Hz requires a time constant of 16 ms, which should mean the differentiated signal getting close to zero in around 50 ms.


Geographer:
The second waveform in the first post would have an amplitude of about 12 V peak to peak at 5 V per division. Is that plausible?

The signal in the first photo looks like it might be coming from something operating at 3 to 3.3 V. Is that plausible?

What sort of probe are you using?

Most analog scopes had provision for variable gain using a small knob somewhere in association with each channel. Be sure that is set to the "off" (normal gain as marked on the main vertical factor knob).

The scope will have a probe calibration output on the front somewhere. The amplitude that is marked probably isn't really accurate, but should be reasonably close. You can also check that your scope cal is OK at low frequency by looking at the power supplies you are using.

I ran across something that gave me the impression that Antaris is from U-blox. Is that true? In any case, do you have a link for a datasheet for the board you are using?
 

Thread Starter

Geographer0912

Joined Nov 19, 2018
7
Yes, U-blox Antaris LEA-4H is the model of the GPS that we were testing our LiDAR with. I have attached the datasheet for it. This was a means to tell us whether or not our LiDAR is in fact working or if the GPS we want to use is not sending the necessary signals. The PPS signal of the first image is from the GPS we are troubleshooting, the second image is the PPS from the Antaris GPS.

Second signal: I am certain I used 5V/div in that measurement, however, whether or not the time pulse output is indeed 12V, I will have to dig further into the documentation for the answer.

First signal: Yes I believe it is plausible, It is a 5V TTL signal.

The probes in use are Tektronix P2200. There is a "Probe Adjust" slot that I have been adjusting the probe compensation with. Similarly, I have used it to make sure the oscilloscope is somewhat calibrated by calculating Hz and reading the voltage.

You can also check that your scope cal is OK at low frequency by looking at the power supplies you are using.
Can you elaborate on what exactly this means?

I really appreciate your help and input.
 

Attachments

Reloadron

Joined Jan 15, 2015
7,501
Can you elaborate on what exactly this means?
On your scope, a TEK 2235 if you look between the Intensity A/B pots and the CH 1 Vertical Position pot you will see a small terminal labeled AMP CAL .5V and a square wave symbol. That signal is a 0.5 Volt peak to peak 1 KHz square wave suitable for checking your probe compensation and vertical gain. It is centered just below those two knobs. Additionally just beside the vertical position knobs, sort of below the knobs you will see two small lamps labeled UNCAL. If you rotate the center red knobs on the CH1 or CH2 VOLTS/DIV knobs those little lamps should illuminate telling you the Volts/Div is no longer in a calibrated state.

Ron
 

ebp

Joined Feb 8, 2018
2,332
I'm sorry. I see how my statement is confusing. I should have said "at DC" instead of "at low frequency."

I just meant that if you use the oscilloscope to check the power supply voltages, which you presumably know or can measure with a meter, you can get a good idea if the scope displaying what you think it should in terms of DC amplitude. There might be a little "fuzz" on the DC level, but you can ignore that. This is nothing more than a simple "confidence" test.

I'll have a look at the U-blox datasheet. I've played with one of their products and it is very impressive.
[EDIT] - datasheet says of timepulse output "1 pulse per second, synchronized at rising edge, pulse length 100ms". There is no diagram of the waveform, but I would interpret this to mean a 100 ms HIGH pulse, so essentially the inverse of your second photo.

Is the GPS module mounted on a secondary circuit board with some other bits and pieces? If so, is there any documentation available for that board?
 
Last edited:

Thread Starter

Geographer0912

Joined Nov 19, 2018
7
Reloadron: I believe I have a slightly different version of the TEK 2235 than you. I do see the small UNCAL lamps you have referenced in pictures of the 2235 on google. However, they do not appear on mine. I've attached a picture of the controls/connectors/indicators of what I'm looking at. AMP CAL .5V on your model appears to be known as Probe Adjust on mine (notated as 500mV p-p, 1kHz)

ebp: I see what you're saying. I just read a 12v power supply with a DMM and the Oscope and they were roughly the same (DMM read +11.87V; while the signal on the oscilloscope was about +2.0 to 2.5 divisions at 5V/div).

Yes, it is part of the SuperSense evaluation kit (AEK-4H). I have attached another document that is relevant to the whole kit.

I will get back to you both once I do some more research on your suggestions.
 

Attachments

ebp

Joined Feb 8, 2018
2,332
It would appear from the AEK summary that the timepulse is directly used to drive a LED, but there is only the block diagram. I can't find anything on the ublox site about the AEK - search, browsing dev & aval boards - nothing! I was hoping to find a schematic.

I'm beginning to wonder if perhaps the problem is what you used for "ground" for the oscilloscope for the measurement. At this point I just can't think of anything probable that would account for the differentiation - somewhere in the circuit the signal is being coupled through a very small capacitance, or in some odd way that makes it appear so. With the probes set for 10:1 attenuation, the low-frequency cutoff of the probe plus AC coupling cap in the scope would be 1 Hz, and clearly from the narrow spike, AC coupling can't account for what is seen.
 

Reloadron

Joined Jan 15, 2015
7,501
Reloadron: I believe I have a slightly different version of the TEK 2235 than you. I do see the small UNCAL lamps you have referenced in pictures of the 2235 on google. However, they do not appear on mine. I've attached a picture of the controls/connectors/indicators of what I'm looking at. AMP CAL .5V on your model appears to be known as Probe Adjust on mine (notated as 500mV p-p, 1kHz)
Well I'll be. They are different and actually mine is a Tek 2235 but it a AN/USM -488 version which is a Mil Spec flavor. They are slightly different and thanks for bringing that to my attention.

Ron
 

Thread Starter

Geographer0912

Joined Nov 19, 2018
7
ebp: You are absolutely correct! The grounding was the issue. We have obtained the expected PPS signal using the "PPS Out" pin grounded by the "PPS Return" pin. Once we connected the GPS and LiDAR's respective wires to one another, we receive error messages surrounding PPS (i.e PPS line is Noisy, PPS Missing, etc) and it failed to connect.

Will physically combining PPS Out and PPS Return pins with a length of wire yield the appropriate signal? I have attached both pinouts for the GPS and LiDAR

If not, can I use something like a multiplexer to combine PPS Out and PPS return while using Pin 15 (Ground) of the GPS for grounding?

The problem is that I have 3 wires on the LiDAR end for NMEA,PPS, and Ground (shared), while it appears I need either:

1) combined PPS signal from PPS Out and PPS Return

2)Split up Ground so that there is a PPS ground and a NMEA ground independent of eachother from the LiDAR

3)Normalize the ground “PPS Out” so that it is the same and can be shared with the NMEA Ground

Again, thank you for your input!
 

Attachments

ebp

Joined Feb 8, 2018
2,332
I wonder if perhaps the PPS output is via an optocoupler, although I can't really rationalize that very well.

Do you have any schematics for the board? Could you post some photographs? (one of the whole board, another one or two of the area around the PPS output of the GPS module itself and whatever appears to be between that and the actual "final" output of the board)

I have a feeling that some small bit of information will take us to a very simple resolution.
 

Thread Starter

Geographer0912

Joined Nov 19, 2018
7
ebp: Sorry about the delay. Hope you had a good Thanksgiving.

Unfortunately, I do not have any schematics for the board but I am trying to get them. All of the components are sealed in a housing that I would regret taking apart.

From reading about differential signals on this site and elsewhere, it appears that they are used for more resistant/stable pieces of hardware, in this case it would be the GPS. Since our LiDAR is perhaps not as advanced and is looking for a single ended PPS signal, is there an "easy" way of turning the differential signal from the GPS into a single ended signal? I get that there isn't a straight forward answer without knowing exactly what's going on the GPS but any recommendations are still appreciated.

So far, it looks like buying/building an amplifier to do this is our best bet.
 

ebp

Joined Feb 8, 2018
2,332
I suspect that all you need to do is tie the return for the PPS signal to "signal ground" but I can't be sure that is safe. I don't think it is a differential signal - I've never encountered the term "return" for such, rather + and - symbols or something along that line. The use of "return" is what makes me suspect an optocoupler - basically a transistor with the emitter connected to PPS Return and the collector to PPS Out.

Exactly how did you have things connected when you discovered that using the PPS Return made PPS Out look right on the scope?
 

ebp

Joined Feb 8, 2018
2,332
Just occurred to me: When you say all of components are in a sealed housing are you referring to what looks like an aluminum box with a USB connector and an RF connector as shown in the PDF at #12, or what is inside of it? Inside is presumably a circuit board nearly as big as the box with another smaller circuit board soldered to it and on the small board the two little sealed ublox modules. It is the parts on the big board that I'm interested in - however if you can get a schematic, that should answer all questions.

And something else - Perhaps I created confusion by referring to the PPS signal as differentiated. "Differentiated" means much as it does in math - the output is proportional to the differential of the signal with respect to time, so you get a positive going spike at the rising edge of the pulse and a negative going spike at the falling edge. A "differential" signal is one where there are two signal conductors and you recover the signal by looking at the arithmetic difference between them e.g. one line might be 2 V and the other 3 V in one logic state and 3 V & 2 V in the other state - it is the +1/-1 volt of difference that counts.

I also noticed mention of
AEK-4T: Adapter board to provide
• 1 UART
• TIMEPULSE,EXTINT0 signals

on the sheet form the PDF at #12.
If such an adapter is involved THAT is probably what we need to sort out. I found some schematics for some other dev kits and they all convert the PPS signal to RS-232 voltage levels (about ±9 volts). It seems very odd to me, but perhaps it is done because you can be really cruel to an RS-232 line driver without breaking it. It may be, however, that if there is an adapter board, it actually takes low voltage signals and makes them into RS-232 levels.
 
Top