Electromagnetic interference with speedometer

[JMD]

Hi all,
I have Dakota Digital's VHX gauges in my '68 k10. The speedometer gets its signal from a transfer case mounted vehicle speed sensor which is also made by Dakota Digital. Everything has been working great for the past 3 years, that is until I ditched the carb and installed a Holley Sniper EFI system.
Now as soon as the truck starts moving, the speedometer jumps all over the place. Sometimes I'll be barely moving and the speedo will jump straight to 100mph. While not having an accurate speed reading is annoying, I'm more concerned that it's more than tripling the miles that should be reading on the odometer.
I know I'm getting electromagnetic interference somewhere in the VSS circuit. I verified this by bringing the truck up to speed, putting it in neutral and turning the ignition off. As soon as the engine and fuel pump stop running, the speedometer reads normally.
So my question is, what part of the Dakota Digital VSS circuit is most susceptible to EMI? The VSS wire is running parallel with the fuel pump power wire, but I made sure to keep them at least 10" apart down the length of the frame. Is that not enough separation? The only other trouble spot I can think of would be interference from the Holley ignition system getting to the brains of the Dakota Digital control box. But again, all ignition wires are at least 10" from the control box. I'm at a loss here, any suggestions?

[Dead Parrot]

A couple of general ideas.
Double check grounds. Did all of the engine to cab/frame ground cables get re-installed? A bad/missing ground might cause current to try to us some of the signal wires as a circuit path.

Run the truck in the dark after your eyes have adjusted and look around all of the ignition wires for any escaping sparks. Possible one of the wires has a defect and is letting some of the fire escape. It doesn't take many 30k volt electrons to mess with computer level signals.

[SkidmoreGarage]

Do you know how many pulses per second the signal is? That may help you understand what could be generating a signal, if it is back feeding through the cluster or sensor.

My gut says to follow the suggestions above, if the ground is loose or shared with a noisy source it could be creating a voltage offset that is lowering the threshold for the trigger to where it is picking up noise. If you have access to a voltmeter, you may be able to clamp onto the signal wires and see if anything funky is happening (voltage increase or decrease) when you see the speedo change. I'd also double check the wire insulation for knicks.

If you believe it is fuel pump related (strikes me as an odd source as it is probably a regular dc motor, but could be related to the pump relay if it is opening and closing), you could temporarily move the speed sensor wire and see if it is resolved.

[JMD]

Thanks for suggestions guys. The speed sensor I have produces 128k pulses per miles. I ran the power wire for the fuel pump down the opposite frame rail and it made no difference, so I guess that can be ruled out. Bummer because that would have been a very simple solution.

The power wires for the fuel pump, gauges, and the ECU are all run directly to the battery as the manufacturer requires. Maybe there’s some interference there? There should be plenty of grounds. Two heavy 1ga wires going from the negative battery post to the engine as well as the frame, and then 4 ground straps running from the body to the frame. The ECU and the gauges ground straight to the battery while the fuel pump grounds to the frame.

I wonder if running the Dakota digital brain box off of a separate battery would narrow down the problem?

[SkidmoreGarage]

That sounds like generally a good wiring strategy, I think the only thing to consider is whether you have multiple grounding ring terminals on the same lug. That can cause problems. Convention is to crimp multiple wires to the same ring terminal or to use separate lugs. But I wouldn't put money on that being the solution.

I made a big assumption that the speedo is a self-scaling analog circuit looking for a square pulse when I was thinking a ground/voltage offset problem. What does the voltage feed look like to the gauge/box/sensor look like? Does it share a circuit with anything?

I would also look at the ECU wiring, does the ECU have VSS in? Are you driving any kind of IACV? I'm not familiar with the sniper efi.

The signal at 30mph is I think around 1kHz, and should be linear. Have you called DD? It may just be a bad box.

[The Rocknrod]

I had some Dakota Digital units. The signal wire from the transmission speed sender. I used a coax type cable, I ground the jacket and never got interference. Takes some doing but was worth it. Good grounds are the key.