Has Anyone Ever Had This Problem With Their Pedals?

[Neilski]

All 3 pedals use loadcells.

Ah, didn't remember that. Depending on how the load-cell amp and ADC are hooked up this would still be compatible with my narrative though

PSU is an Antec HCP-1200. HCP-1200 is very high-end with OEM Delta parts inside. It's comparable to PSU series like Corsair's top-of-the-line AX and EVGA's G2/P2/T2.

Damn, plenty of grunt. Maybe it's just faulty then. (But see below, cos it shouldn't truly matter to any devices.)

How much GPU power was used with your stress test?

Between CPU and GPU, well over half of the rated power of my 430 W supply. (Possibly a bigger fraction than you were pulling?)

Multimeter and powered USB hub arrived. Will test in the next hr or so.

Multimeter is probably not needed now that you have clear evidence of voltage droop, but the powered hub will be fun.

If any drop or rise in 5v rail results in different pedal readings, would you say that's a pedal design flaw (or at least design oversight)? Even of the 5v rail doesn't drop to 4.9-ish V, it'll still drop nonetheless with any PSU (eg. 5.13 V to 5.05 V).

Yes, I'd call it a clear flaw. I reckon the pedals shouldn't depend on that voltage being steady (or a 1% voltage variation would make anything more than an 7-bit ADC pretty pointless), and it seems likely that they do. If the powered hub fixes the issue I guess we can feel certain of that.

(When I said my supply was rock solid, I meant even less variation than your example - the total range on the +5 V was only 40 mV, with no correlation at all between the voltage and the load and nothing in between 5.04 and 5.08 V; it just flickered between the two now and then, but sat mostly at 5.08. I speculate that (a) my PSU has no coupling between the 12 V and 5 V modules and (b) the voltage readout on my motherboard has insufficient resolution to see (say) 5.06 V... So to be totally fair to your pedals, if my G27 had the same flaw I wouldn't actually notice it given that my +5V line is stable.)

 

[Spinelli]

Tried with the powered USB hub (tests repeated around 10 times).

Desktop, low load, +5V rail @ 5.02 V, 100% physical throttle, firmware raw value:

• motherboard USB = -9000
• powered USB = -8200

Keep in mind above that the +5V rail reading is the same (5.02 V) yet the direct-to-PC method is still showing a drop relative to the powered USB hub.

ACC with high load (gfx settings in other post), GPU @ 545 W, CPU @ 145 W, +5V rail @ 4.9 V, 100% physical throttle, firmware raw value:

• motherboard USB = -9700
• powered USB = -8200

Conclusions:

1. Even at low load with the same +5V rail reading, the motherboard USB is still reporting a drop in pedal reading compared to USB powered hub
2. USB powered hub solves the issue. Also tested with clutch, readings are identical (15,300) between desktop low load and ACC high load

Questions:

• Due to conclusion "1", does that mean this issue is not purely related, or possibly at all related, to the +5V rail?
• How many others are using pedals where differences in load, or even just the difference between using a direct-to-PC connection and a powered USB hub connection - regardless of load - results in different pedal input values without the user even realising?
• Does this mean I have to make sure nothing else is connected to my 24 W powered USB hub - even if the devices use like 1 W - since there's clearly some weird over-sensitivity happening with the pedals and power loads?

 

[Deleted member 197115]

There is a good section on the wiki covering powered usb-hubs. In general properly designed self-powered USB hub should be able to supply 5V/500mA to all ports, unlike bus powered.

 

[Neilski]

USB powered hub solves the issue.

Yay!

Due to conclusion "1", does that mean this issue is not purely related, or possibly at all related, to the +5V rail?

No I think it probably doesn't mean that. But you can use a multimeter to put it beyond doubt of course.
Extrapolating from your measurements, I could guess that the powered hub is simply delivering a higher voltage to the pedals - a tad over 5.1 V. This is not at all unusual for a powered hub. (USB spec has a few different tolerance definitions, but the tightest of those is 4.75 - 5.25 V so it's well within the spec.)

Does this mean I have to make sure nothing else is connected to my 24 W powered USB hub - even if the devices use like 1 W - since there's clearly some weird over-sensitivity happening with the pedals and power loads?

Depends on how good your hub is at regulating the voltage. Only testing will tell you, but you can always manually add a bit of dead zone at the top of each range if the variation isn't crazy big under extra hub load.

To reiterate: I reckon the pedals shouldn't behave the way they do when the voltage sags, but then the voltage shouldn't sag that much as the PC load increases. I imagine that almost everyone who bought the pedals has a USB voltage which doesn't change much and the precise value matters not a jot as long as it's stable - so once you calibrate the pedals they would stay calibrated.

 

[Spinelli]

Yes, I think there might be a design oversight because the 100% throttle input drops even in a low graphics load situation like, for eg., RF1. It doesn't drop as low as the high load ACC gameplay but it still drops compared to desktop. It's like an increase in load results in a corresponding decrease in input no matter how big or small.

I have a feeling if this was common with other pedals, there'd definitely be other ppl mentioning it, especially people that are very picky & analytical.

Like you said, it could be something in the pedals' electronics that is way too sensitivity to voltage variations - even tiny ones that'll happen on all systems.

Now I don't know if I should get a new PSU. My PSU's inner parts are very high end and it powers overclocked and overvolted high-powered CPUs GPUs (eg. 4090 at the full power), RAM, etc
all day, without a hitch.

 

[Neilski]

Now I don't know if I should get a new PSU. My PSU's inner parts are very high end and it powers overclocked and overvolted high-powered CPUs GPUs (eg. 4090 at the full power), RAM, etc
all day, without a hitch.

Well, I can't say I'm certain that the PSU has total responsibility for the USB voltage (but it's what I'd guess). Perhaps the motherboard also plays a role...?
Someone else may know, or Google may throw up an instant answer...

 

[Spinelli]

Well, I can't say I'm certain that the PSU has total responsibility for the USB voltage (but it's what I'd guess). Perhaps the motherboard also plays a role...?
Someone else may know, or Google may throw up an instant answer...

Ya. It's strange. I'm glad it's resolved with the powered USB but not satisfied because I still don't know why it happens.

The PSU voltage drop may be a bit more than normal but it's certainly nothing extreme.

Also, there's a pedal input drop even during low PC loads (eg. rFactor 1). The calibration input drop is much less at low load than high load but it exists nonetheless. This makes me feel like there's something with the pedals that go up and down in accordance with some sort of electrical/power aspect of the PC and therefore it's not, or at least not entirely, a PC/PSU issue. You mentioned before that this sort of behaviour could have something to do with the design of the pedals with regards to the ADCs, amps, loadcells, etc. rather than the PC/PSU (or purely the PC/PSU). Is there a fairly simple way to determine this for someone who isn't too electrically inclined?

 

[stigs2cousin]

I personaly (as a mechanically gifted and electronically challenged person) can only think of using the LeoBodnar load cell board.
If with that you still have the problem I´d call Olivia Dunham (Fringe Division)

But that would be "throwing good money after bad"

Either be happy that it works moderately well or send them back.
(german proverb says: "buy cheap, buy twice" )

 

[Neilski]

Also, there's a pedal input drop even during low PC loads (eg. rFactor 1). The calibration input drop is much less at low load than high load but it exists nonetheless. This makes me feel like there's something with the pedals that go up and down in accordance with some sort of electrical/power aspect of the PC and therefore it's not, or at least not entirely, a PC/PSU issue. You mentioned before that this sort of behaviour could have something to do with the design of the pedals with regards to the ADCs, amps, loadcells, etc. rather than the PC/PSU (or purely the PC/PSU). Is there a fairly simple way to determine this for someone who isn't too electrically inclined?

I don't think I'm understanding your question here, as I had thought the issue was basically already put to bed.

Here's the logic in one place for assumption checking:

HWinfo indicates that the voltage on the +5V line drops with system load. I have proposed that this probably affects the voltage fed to USB devices. (A multimeter would support or invalidate this assumption, and this may be straightforward to test, but if you aren't confident about doing the test then I wouldn't strongly recommend it.)

Following on from this assumption, a guess about why the pedals behave in a load-sensitive way would be that they are using the USB voltage to drive the load cell (producing an imbalance voltage which will be proportional to the USB voltage), while using an ADC voltage reference which is not derived from the USB voltage. This will cause the digitised pedal reading to vary with the USB voltage, at any point when the load cell Wheatstone bridge is imbalanced (this could be any pedal position but likely to be when you push down on the pedals).

At this point I reckon the simple way to determine what's going on is to just ask the manufacturer if their electronics design has this behaviour

 

[Deleted member 197115]

Olivia Dunham (Fringe Division)

Love that show.

 

[Neilski]

Love that show.

Ah, hadn't heard of it but looks like it might be fun

 

[Spinelli]

I don't think I'm understanding your question here, as I had thought the issue was basically already put to bed.

Here's the logic in one place for assumption checking:

HWinfo indicates that the voltage on the +5V line drops with system load. I have proposed that this probably affects the voltage fed to USB devices. (A multimeter would support or invalidate this assumption, and this may be straightforward to test, but if you aren't confident about doing the test then I wouldn't strongly recommend it.)

Following on from this assumption, a guess about why the pedals behave in a load-sensitive way would be that they are using the USB voltage to drive the load cell (producing an imbalance voltage which will be proportional to the USB voltage), while using an ADC voltage reference which is not derived from the USB voltage. This will cause the digitised pedal reading to vary with the USB voltage, at any point when the load cell Wheatstone bridge is imbalanced (this could be any pedal position but likely to be when you push down on the pedals).

At this point I reckon the simple way to determine what's going on is to just ask the manufacturer if their electronics design has this behaviour

Yes, I'm trying to find if the problem is the design choice of the pedals being directly reliant on the PC's +5 V in such an extreme and sensitive fashion that the pedal input (for any particular physical pedal position, eg. 100% throttle throw) will continuously change since the detected input goes up and down in direct relation to system load, even if the +5 V rail goes from, let's say something like, 5.006 V to 5.005 V and probably micro-changes that aren't even measurable with HWinfo. These are system loads and voltage drops that are completely normal. Even a brand-new, good PSUs will have voltage drops on the +5 V rail of 0.3 to 1.x % percent.

In other words, the voltage drop between idling at the desktop and playing RF1 with a capped framerate and no AI is so small, it's not even recognized by HWinfo, yet even with such a light load, the input value still changes compared to idle desktop (just no where near as much as during a heavy load). So that seems like a pedal design over-sight while the ACC voltage drop - while bigger than ideal - isn't the real cause.

EDIT: You know what? They've been perfect - consistently perfect - with the powered USB hub so, forget it, I'm not going to pursue finding out exactly what's going on electrically. I'm too busy having fun with them.

Either be happy that it works moderately well or send them back.
(german proverb says: "buy cheap, buy twice" )

With the powered USB, it doesn't just work moderately well, it works absolutely perfect

Been playing 3 or 4 days now, the values of the pedals are still identical. No change in 0% physical input value or 100% physical value no matter if the PC is idling or pushing 1000 W, if the PC restarts, is turned off for hours, if the pedals are getting beaten, if they're disconnected, etc. It's completely resolved - absolutely perfect - with the powered USB hub.

 

[stigs2cousin]

With the powered USB, it doesn't just work moderately well, it works absolutely perfect

Finally.

Took a lot of people a lot of time, but we got there.

 

[Spinelli]

Finally.

Took a lot of people a lot of time, but we got there.

I'd like to thank everyone who posted. I really appreciate all the info, ideas, opinions, etc.