4425A glitch on HV ?

Ask for and share advice on using the PicoScope kit to fix Electric Vehicles
Post Reply
picowight
TwoWaves
TwoWaves
Posts: 119
Joined: Sat Oct 26, 2019 6:45 pm

4425A glitch on HV ?

Post by picowight »

Alphard Hybrid 1st gen - no crank or half a crank
This has been round the houses with I think techs only looking at the scan tool data of the HV ECU that would say modules (2 cells) at 14.4v (nominal 7.2v) then later saying some cells were at 0v and others at 4v, then saying everything was fine again. HV ECU was swapped out and didn't fix it, battery charged with special charger and didn't fix it, a suspect cell was changed and didn't fix it, then they were talking about changing out the £5K inverter as there was a code for that... Then it gets to me.
I look at the actual voltage both sides of the relays during cranking and it just tanks. Red is Battery side of relays blue is load side - this is right at the battery. Measuring separate grounds at the relays.
Screenshot from 2022-09-15 22-50-28.png
What I find odd is the glitch at 6.2 s where both voltages disappear. The blue is on a 20:1 attenuator, red (battery) is direct.
Measuring just the output side voltage and the HV current, that drop to beyond -200v is not there - is this an issue with the floating earths ?
Screenshot from 2022-09-15 23-14-27.png
Then measuring both again along with the 12v to the HV ECU and the current it is back. This time there is no crank it looks like the relays are not fully closing and I am just seeing the a v drop across the relay resistor I think (no separate pre-charge relay that I can see - its a Japanese import and pre HSD so no wiring diagrams) as batt voltage stays high and there is not enough juice to even half crank the engine so the relays trip in and out.
Screenshot from 2022-09-18 12-26-15.png
Conclusion is that the battery is dead - showing voltage but not holding any charge - I did individual drop tests on modules with a 3A test light and was seeing a significant v drop.
But really my question is about that "gap" - Here is the same measurement but with the blue and red on the HV battery ground on the input side of the ground relay. - no glitch
Screenshot from 2022-09-18 14-38-48.png
All of the files are 30mb + too big to upload here so here is a dropbox link
https://www.dropbox.com/sh/kyyv9ggs3kr5 ... KzrBa?dl=0

Steve Smith
Pico Staff Member
Pico Staff Member
Posts: 1356
Joined: Sun Aug 25, 2013 7:22 am

Re: 4425A glitch on HV ?

Post by Steve Smith »

Hello and thank you for the post and taking the time to share this adventure

Regarding “But really my question is about that gap

Could you confirm during the glitch at 6.2 s (where no data is drawn) if the scope momentarily displayed a Common Mode Range Error, denoted by an exclamation mark inside a yellow triangle?

A little more info can be found here topic17121.html and here https://www.picoauto.com/download/docum ... de.web.pdf

In image 1 (Capture 20220915-0005 hv on crank) you mentioned “Measuring separate grounds at the relays” could I clarify how channels A and B were referenced? i.e., to vehicle chassis ground or HV ground?

The reason I ask is because the waveforms display no data during the cranking/power up period which suggests a possible Common Mode Range Error. Here, the voltage present on the test lead ground wires (Channel A or Channel B) has exceeded +- 30 V from nominal oscilloscope ground (Typically 0 V chassis ground)
Image 1
Image 1
Looking at the following capture, image 2 (Capture 20220915-0008 v a crank 2) does not demonstrate this symptom as channel A captures the increase in voltage with no “gap” Note: Channel B is not used, suggesting our common mode range error maybe linked to channel B ground point?

Image 3 (Capture 20220915-0013 v2) displays loss of data as Channel B is once again active. Note how loss of data applies to channels utilizing ground references. (Channels A, B & D)

I will assume Channel D was referenced to chassis ground?

Channel C looks to be a current clamp which maybe utilizing the ground reference of an internal battery and therefore unaffected by data loss once the Common Mode Range error was removed post cranking /power up (please forgive any misinterpretation or assumptions)

Finally, image 4 (Capture 20220915-0010 common ground pre relay) has no data loss now Channel B is referenced to HV Negative suggesting the Common Mode Range Error is no longer sufficient to be displayed; i.e., below +- 30 V from nominal oscilloscope ground (Typically 0 V chassis ground)

Over and above the data loss and “gap” you describe I have a far greater concern and that is the probes used to measure High Voltage capable of delivering high energy

Whilst the PicoScope 4425A can handle a 200 V input comfortably, Channel A utilises a x20 attenuator and Channel B a x1 test lead, neither of which offer the protection required for high energy, high voltages exceeding 60 V DC

The x20 attenuator, while typically rated at 300 to 400 V, these values apply to low energy sources such as primary ignition spikes where voltages are only at these levels for microseconds (µs). In addition, only the signal terminal of the attenuator is attenuated and therefore, should the ground of the attenuator be inadvertently connected to a high voltage / high energy point on the vehicle, the risk of Common Mode Range Error and possible electric shock is increased

Moving onto the x1 test lead, again this lead will be sufficiently insulated to protect the user, however it lacks attenuation and therefore the required fail-safe impedance for such measurements

In both cases above (x20 attenuator and x1 test lead) the answer is to use the active differential probe https://www.picoauto.com/products//elec ... e-x20-x200 which incorporates the relevant impedance for both the ground and signal wires at your test point along with the appropriate attenuation to ensure safe working voltages (signal and ground) at the scope at all times

Should the vehicle then be connected to the grid (mains) if applicable (e.g., PHEV) then neither the x20 attenuator or the x1 test lead include the relevant CAT Rating

I hope this helps, take care……..Steve

picowight
TwoWaves
TwoWaves
Posts: 119
Joined: Sat Oct 26, 2019 6:45 pm

Re: 4425A glitch on HV ?

Post by picowight »

Lot to take in there. Thanks Steve.
Yes there was a flash of something yellow but too quick to read.

A was the output side of the relays on + and -ve relays, B was battery side of those relays. So both grounds on the same HV -ve but on either side of the relay. From your explanation I am assuming the issue was the closing time/voltage ramp of the -ve relay - since as you point out it's not there when using a common connection point on the battery side of the relay. So the "gap" was the scope protecting itself - which is what I suspected hence asking the question.

Yes D was measuring the 12v supply to the HV ECU so referenced to 12v battery/Chassis ground. Correct C was a current clamp

The voltage clamps used were CAT I / 1000v rated as were the gloves.
Thanks for the explanation. In future I will avoid measuring more than one switched ground on HV systems and buy some attenuated scope leads the Pico TA400 does not seem to be available so something like the RS 10:1 https://uk.rs-online.com/web/p/oscillos ... es/1799564 as a safer way of attenuating HV sources.

EDIT - found it on the very odd product menu, https://www.picoauto.com/products/test- ... be-adaptor

Post Reply