Part jumps up on vacuum release (no delays, no blow, all coordination = None)

[Artem Stanchak]

Hi everyone,

I’m running into a part “jumping up” issue during placement, and I’d like to confirm the correct / recommended behavior in newer OpenPnP versions.

Symptom:
When the vacuum is released, the part does not just detach — it actually jumps upward for a moment before Z lift.
This causes unstable placement on small parts.

Important details:

All vacuum-related actuators:

Before Actuation = None

After Actuation = None

Before Read = None

No vacuum on/off delays configured

No blow-off / air assist enabled

Same hardware worked correctly with an older OpenPnP version

Observation:
It looks like:

Vacuum OFF and Z lift happen too close together

Residual vacuum + fast Z acceleration causes an air rush into the nozzle

This creates a “syringe effect” that pops the part up

Questions:

Is this behavior expected with the current placement sequencing?

Is there a recommended way to:

add a small dwell (e.g. 5–15 ms) after vacuum OFF?

or slow down Z acceleration only for placement lift?

Is there an existing option, pipeline stage, or config setting for a soft vacuum release?

Has placement timing or Z lift behavior changed compared to older OpenPnP versions?

Any insight or best practice would be greatly appreciated.

Thanks!

[Toby Dickenson]

Hi Artem,

> add a small dwell (e.g. 5–15 ms) after vacuum OFF?

The placement dwell time occurs between changing the vacuum actuator
and lifting the nozzle, so you already have this. I have no blow-off
and dwell times up to 100ms can be needed sometimes. 15ms is a tiny
dwell time.

> Residual vacuum + fast Z acceleration causes an air rush into the nozzle

How do you get "residual vacuum"? disabling the vacuum valve on
placement should vent the nozzle to atmospheric pressure (if you have
no blow-off)

Toby

[Jan]

Hi Artem!
"Before Actuation" shall be enabled on the vacuum valve actuators. It
enforces that the machine has stopped before the valve is operated. The
default behavior of G-Code controllers is to queue motion segments (eg.
G1) and execute M-Codes immediately. This might under worst case
conditions result in the vacuum being switched off before the nozzle
arrives at the place location.
In addition a place dwell time can be configured per Nozzle and Nozzle
Tip (sum of both are used). Please note, that the delay is executed
after the vacuum valve has been switched off. Without "Before Actuation"
this can happen while the machine is still moving.
Please also note, that dwell delays can be offloaded to the controller.
You need to define "DELAY_COMMAND" G-Code for that. ("Before Actuation"
is still required.)

Jan

> --
> You received this message because you are subscribed to the Google
> Groups "OpenPnP" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to openpnp+u...@googlegroups.com
> <mailto:openpnp+u...@googlegroups.com>.
> To view this discussion visit https://groups.google.com/d/msgid/openpnp/
> bb6e7690-881c-4915-97a5-76512334aaa5n%40googlegroups.com <https://
> groups.google.com/d/msgid/openpnp/
> bb6e7690-881c-4915-97a5-76512334aaa5n%40googlegroups.com?
> utm_medium=email&utm_source=footer>.

[vespaman]

Hi Artem,

add a small dwell (e.g. 5–15 ms) after vacuum OFF?

Unless you have a double actuated valve (two coils), that is way to little! A pretty fast single coil valve will take about 10ms before reacting, an normal valve will take longer (the actuate off is by spring, so much slower than actuate on).
There are superfast valves (expensive ones), that might reach down to/below 5ms. (I have been drooling, yes!)

But then the vacuum has to be evacuated. My valve has 10ms on release, then about 13ms before the vacuum is mostly gone (on a 502 (0402-type) nozzle tip). This is with the valves on the head. If your valve is on the floor of the machine, the evacuation will take longer time.
So it will depend on your valve, valve placement, your tip, general leakage of course and the vacuum level you are starting with.

Increase that dwell time! 

 - Micael

[Artem Stanchak]

Thanks everyone for the reply. I will use a delay on the nozzles.
Yes, my valves are on the bottom of the machine.

пʼятниця, 16 січня 2026 р. о 00:25:52 UTC+2 micael....@gmail.com пише: