I've been using generated wav files for bench trigger testing for a long time -with success.
Now I managed to have an Asus Eee 900 laptop for tuning, and it seems it has a different soundcard. If I attach the ECU to the headphone output, the ECU jumps to fine rpm for 2-3secs then it loses rpm signal (sometimes with trigger error).
It is the case every time I disconnect and reconnect the audio cable. -I need to disconnect, it's not enough to stop and restart playing from software.
Other notebooks work flawlessly in the very same situation and cables, but this tiny one (eee 900) not.
It's sound is fine listening through an earphone but not for VEMS
Any ideas?
- after loss signal vems is not receive anything? neither after vems reset? Maybe turning of device detection on soundcard out and in can help...
- VEMS power down and up helps also.
I've checked the signal with a scope, and the problem is that vems seems to pull-up the whole signal.
It is normally a square signal between -2V and +2V, but as VEMS is connected (or powered up), its climbing and in a few secs it ending up just below zero (range of approx. 0.5V and 4.5V), since the essence of this signal is crossing zero, now I see why it looses rpm.
I guess this soundcard is isolated with some series capacitor inside which allows some DC bias of the signal(?)
Question: how to solve this problem, -preferrably outside of both vems and the notebook?
2k7 pulldown recommended on trigger input when driven from soundcard
I had the same problem, some sound cards were working, some not. Every sound card I came across has a DC-blocking cap. The VEMS has a pullup resistor making a DC-offset of 5V : (2k7 pullup by default for HALL-type trigger)
- VR usually works fine with every soundcard, but a 2k7 .. 10k pulldown does not hurt anyway
- Note: VR used to have 18k pullup, has 27k pullup since about 2012. The clamping diodes will make the peak value 5v, thus making the DC offset effectively smaller, but if the amplitude of your sound card is to small, it will not reach the '0' level.
The solution is very simple: Connect a pull-down resistor of the same size as the pull-up resistor in your test cable, thus a resistor from the HALL-input(s) to ground of 2k7 .. 10k. This worked perfect on my EeePC 900 and other laptops that weren't working before.
--Tomvl
- Got it! 10kOhm pull-down resistor to GND worked! (100k not), thanks for your help!
- 2k7 pulldown is the most universal solution (best for VR or HALL or wheelspeed inputs when driven from soundcard).
- For VR, the pullup usually has no adverse effect if left there (minor attenuation of signal amplitude, and noise also).
- From HALL or wheelspeed inputs: remove the pulldown when driven from real sensor