Connecting Heusinkveld Ultimates to the Fanatec CSPv3 Controller Board

[Atak_Kat]

Here is the result of the testing with the load cell. Few comments:
- same method as above testing/connections done for hall sensor
- I devised a way to mount the load cell assembly as it was on the pedal, and then mount a clamp to the entire assembly for pressure. So it should be giving readouts as intended.
- I made several readings across the pins. Hopefully the chart is understandable what I did.
- In each intersection is three readings. Each representing the pressure on the load cell, as read in the Fanatec Control Panel app (I'm pretty sure I had it previously calibrated quite heavy). I simply could not put more than about 85% pressure with the clamp setup I cobbled together (clamp not heavy enough, setup not so stable. But it was quite a lot of clamping pressure anyway)


Am I correct?
- RED = Excitation +
- BLK = Excitation -
- WHT = Output -
- GRN = Output +

 

[GeekyDeaks]

Am I correct?
- RED = Excitation +
- BLK = Excitation -
- WHT = Output -
- GRN = Output +

I totally concur with that. That is a great bit of detailed analysis!

Right, now the fun begins!

If I was doing this, then my next step would be to make a simple breakout for the connector on a HE pedal and also possibly a breakout for the socket on the CSP board. This should allow fast prototyping of the next stages and let you take the supply voltage from the CSP board to power any amps and provide the Excitation voltage to the load cell, but you might want to consider using batteries whilst messing around to avoid accidentally shorting anything out.

This is the kind of thing I am thinking of, but you can fashion up your own with some sockets and plugs and some jumpers.

You just want something to allow you to directly wire it to the CSP board (for the brake) or a load cell amp circuit. On the subject of the circuit, I see leo bodnar do one, but it doesn't allow you to adjust the gain. This might be a problem as ideally you'd want to be able to adjust the gain so you can get the output from the amplifier in the same range as that from the hall sensor circuit. I'll do a bit more digging.

 

[Atak_Kat]

'Someday I will need that'.... finally has arrived
Digging in a box of old crap that I keep around just-in case, I found some old electronics/phones with various connectors and ports to disassemble.

These should all be RJ10 (I think that's the one that HE uses, which is quite narrow, and has space for only 4 pins/wires.

I have to keep digging to see if I can find RJ12 cable (I think that's the one with 6pins, and all 6 wires). The CSPv3 board uses the outermost pins and I think most of the cables I have only have 4 wires with the outermost pins not connected (RJ11?)

Unbelievably, I don't own a bread board. Probably time to get one. Will make all this much easier. I have several jumper wires from previous Arduino adventures.

If I understand correct, the idea to use battery is likely a good suggestion. If I short the CSPv3 board, then it's game over. Or, I have a soldering station with power supply that can be controlled for voltage, amps, etc, etc. Maybe that's an option to replicate the same voltage?

 

[GeekyDeaks]

Or, I have a soldering station with power supply that can be controlled for voltage, amps, etc, etc. Maybe that's an option to replicate the same voltage?

Yeah, that should work. I have an old PC and Xbox PSUs that I use, but they only give out 12v and 5v so I also have some 3.3v regulators I can use on the breadboard.

I'm struggling to find an off-the-shelf load cell amp circuit that allows you to adjust the gain. You up for a bit of circuit design? Most of the amps only require a small number of passive components to operate. You have some choices for amps, I guess it may depend on what you can source. My G29 mod used the INA122P https://github.com/GeekyDeaks/g29-load-cell#amplifier-circuit. The CSP main board is using the MCP6N11 (https://www.mouser.co.uk/datasheet/2/268/25073A-75155.pdf). Interestingly, the hall sensor is also a wheatstone bridge, so that also has the MCP6N11 amp, the only problem is that I think that amp only comes in surface mount packaging.

 

[Atak_Kat]

Here is some additional info about the HE pedals:

1. HE replied to me, and they advised that the Ultimate+ uses:
- the Brake uses Mavin NA151 (200kg)
- as I understood, the Throttle and Clutch both use Mavin NA27.
- Note that I have the original Ultimate, later upgraded to the Ultimate+ with their upgrade kit, but I'm going to assume that the load cells are the same.

2. HE also confirmed that the pinout wiring is:
Black: - Power supply
Green: + Signal
White: - Signal
Red: + Power supply
(But see below test result, because now I'm a bit confused and I wonder if the +/- here are backwards....)

3. I did some similar testing as done for the CSPv3 load cell:
- I built a small contraption with my plugs/sockets/bits to put in between the pedal and the Ultimate+ Smart Control box, so I could attach the meter leads to the wires easier. No picture since it involved hot-glue and poor soldering, but it worked and I was careful to mark the wiring colours of each of the 4x wires according to those coming out of the HE pedals.
- For each test, I had the Heusinkveld Start Control software open on the PC, and each pedal was tested independently connected to its appropriate port written on the Smart Control box (ie. throttle pedal connected to the throttle input on the box, etc, etc)

Here's the result, with a few comments:
- am I correct that the +/- mentioned above is backwards?
- Note that here, the voltage seems to be 5.1v, where the CSPv3 was only using 3.27v. Not sure it that's another hurdle, or not. I have a feeling that will be a problem, if these Heusinkveld load cells need a higher voltage to function correctly?
- Also, for all of these, the readings are at rest. The only combination I tested with different pedal pressures was the green/white.
- I noticed that in many cases the readings across the same wires, but different pedals, was different. Example the black to green/white, or the red to green/white (between +/-2.33v and +/-2.77v). I sorta assumed they would be the same for each pedal. Seems not.
- For the green/white, I took a reading when resting, and then also near 100%. For the throttle/clutch that was no issue, but for the brake that was quite tough to max out it was likely around 85% for the Brake. Thankfully my little multi-meter has a setting to capture 'max' so that helped to do the brake hands-free.
- Just a note that even when resting, both the Throttle and Clutch were passing some mV. Could be just the way I have them configured with some preload, I assume.


Next small adventure will be to make a test cable to connect the HE pedal output, to the CSPv3 brake input. But I'm gonna wait for a reply here, due to the voltage readings different between the two, and maybe that would be a mistake....

 

[Atak_Kat]

You up for a bit of circuit design?

Sure, why not. Might require some hand-holding but that's why I started the thread!

Have a read my recent post on the Heusinkveld load cells. Might be something important there.
I just ordered a bread board, and took a punt to order 2x of these. Perhaps completely wrong, but they might be useful even to experiment what the resulting output is (before connecting to the CSPv3 inputs)....

 

[GeekyDeaks]

1. HE replied to me, and they advised that the Ultimate+ uses:
- the Brake uses Mavin NA151
- as I understood, the Throttle and Clutch both use Mavin NA27.
- Note that I have the original Ultimate, later upgraded to the Ultimate+ with their upgrade kit, but I'm going to assume that the load cells are the same.

That company never ceases to amaze me.

I'd recommend looking up the details for each. I did a quick check and confirmed my suspicions that they are simply rated for a different range of capacity (load).

Note that here, the voltage seems to be 5.1v, where the CSPv3 was only using 3.27v. Not sure it that's another hurdle, or not. I have a feeling that will be a problem, if these Heusinkveld load cells need a higher voltage to function correctly?

I am probably not explaining this very well, but a load cell is nothing more than 4 resistors arranged in wheatsone bridge. You can do this project without understanding what that does, but honestly, I'd really really recommend you research it first as it will make the process a lot lot easier. This means that they are just passive device and don't require a specific voltage to operate, they will just output a voltage as a % of the input.

Next small adventure will be to make a test cable to connect the HE pedal output, to the CSPv3 brake input. But I'm gonna wait for a reply here, due to the voltage readings different between the two, and maybe that would be a mistake....

As per above, don't worry about it, they are just a bunch or resistors and mV output is proportional to the supply. The important bit is you have the E+/E- and S+/S- identified on both sides. If you get the +/- wrong all that will happen is the signal will be reversed and go down under load, so just switch them around.

 

[GeekyDeaks]

I just ordered a bread board, and took a punt to order 2x of these.

Ah, so that device has a serial output, which unfortunately, is not going to be something you can directly use with the CSP board.

 

[Atak_Kat]

I am probably not explaining this very well, but a load cell is nothing more than 4 resistors arranged in wheatsone bridge. You can do this project without understanding what that does, but honestly, I'd really really recommend you research it first as it will make the process a lot lot easier. This means that they are just passive device and don't require a specific voltage to operate, they will just output a voltage as a % of the input

I didn't comment earlier, but I did read up on it (thank you for the links), and I think I reasonably understood the concept. There are resistors attached/arranged in a special configuration on the material, between the input, and the output. They are essentially flexible, can deform, and put in specific places around where the material will deform when under pressure. As a force is applied, and the material deforms, the resistors also deform (some compressing, some under tension) and this changes how much they affect the electrical signal passing through them. So with a constant input, the output can change depending on the amount of deformation caused in the material. And that output is very small, but very precise. Probably way wrong, but that's how I have it in my head.
What I was unsure of, is that various load cell specs seem to have a 'recommended excitation'. Perhaps a range where the result is most accurate? I think the Mavin 151 is 5-12VDC. But if it's not a critical issue with lower V, then we read 'recommended' somehow similar to the speed-limits driving in Italy. .

The important bit is you have the E+/E- and S+/S- identified on both sides

I'm going to re-test the HE pedals result. It just seems strange to me that the resulting +/- are reversed vs. what expected. It annoys me and now I question if I had it correct.

Ah, so that device has a serial output, which unfortunately, is not going to be something you can directly use with the CSP board.

I figured it might be like that. Oh well, something for a future Arduino project then.

 

[GeekyDeaks]

I'm just looking at the figures you supplied from the HE probing.

Taking just the throttle for now, I can see you measured the following between the S- and S+

resting: 3.6mV
85-100%: 5.9mV

The specifications for the NA27 are here: https://www.mavin.cn/kitchen-scale-load-cell-low-capacity-weight-sensor-na27_p28.html

The two things that are important for us are:

Capacities: 0.6-10kg
Rated Output: 1.0 mV/V

The capacity essentially tells you what load the cell can reliably measure. Too small and it's not accurate, too much and it might actually permanently deform the material. There is alway a bit of tolerance in this, especially on the maximum.

The rated output tells us what to expect at maximum rated load based on the Excitation voltage. In this case we get 1mv per Excitation V. Since we have 5V Excitation, we should expect 5mV maximum out at full load. You are seeing a bit more than that, so I suspect it's over the rated maximum of the cell, but it's not something I'd be concerned about. The fact that it doesn't go back to 0mV at rest probably means it's already under some sort of pre-load.

I'll put together a spreadsheet estimating what the mV outputs are likely to be from each pedal if the Excitation is 3.3V.

 

[GeekyDeaks]

So with a constant input, the output can change depending on the amount of deformation caused in the material. And that output is very small, but very precise. Probably way wrong, but that's how I have it in my head.

That is exactly how it works!

What I was unsure of, is that various load cell specs seem to have a 'recommended excitation'. Perhaps a range where the result is most accurate? I think the Mavin 151 is 5-12VDC. But if it's not a critical issue with lower V, then we read 'recommended' somehow similar to the speed-limits driving in Italy. .

Yeah, I think that is also correct. It's just the range in which the device is guaranteed by the manufacturer to work with the accuracy they advertise. When you using such things in a precision measurement device it's important. Not so much in this application

 

[GeekyDeaks]

I did a quick calc of what I estimate the HE pedals will be at 3.3V

	input V	resting mV	85-100% mV
HE Throttle	3.3	2.376	3.894
HE Brake	3.3	0.132	3.564
HE Clutch	3.3	0.792	3.366

The Brake might be a bit too much outside the range expected by the CSP board since you measured 1.3mV - 4.2mV, but it's still worth a punt as none of this stuff is precise so most manufacturers design a bit of 'self calibration' into their kit, looking for the minimum and maximum values and adjusting as required

 

[Atak_Kat]

The Brake might be a bit too much outside the range expected by the CSP board since you measured 1.3mV - 4.2mV, but it's still worth a punt

Just to note that on the CSPv3 brake I took the first measurement when there was 10% pressure showing in the Fanatec Control panel. While the HE was without pressure. Would it be helpful if I quickly re-tested the Fanatec one while resting? (let's be honest, kinda having fun with this, so I'm gonna do it anyway.....)

Also for the HE measurements, I am sure that I have the throttle travel stop at about 75-80%. So the readings I did won't really give the full range possible. So, just because, I will retest that at full range as well. I also had a closer look at the way the throttle/clutch assembly works and there is a small pre-load on the load-cell, even when resting, just by design. It's not really adjustable and just a result of a little spring in the mechanism. But by getting the full range of the throttle, I think we can say the clutch will have the same range (even if the starting/resting voltage might be slightly different)

I don't think I will be able to make an adapter cable (from HE to CSPv3 brake input) until tomorrow. Had to order some RJ12 extensions with 6 wires and should only arrive tomorrow to be chopped up and sacrificed.

 

[Atak_Kat]

OK, so here is the updated testing.
First the Fanatec Load Cell
- differences vs. prior version highlighted in orange.
- in this case I took a measurement when fully resting/unloading (ie. 0%)
- I was able to get a reading with even higher pressure (95% as shown in the Fanatec Control Panel app), but that's just because I had to clamp the assembly to be able to read from the pins. In reality, it's not so difficult to max-out the pressure when seated and using your leg. Far, far easier to max it out than the Heusinkveld is.
- So it looks like the signal range is quite a bit wider than before?

Next the HE Pedals:
- differences from prior version highlighted (although I didn't highlight it, I also re-checked the resting readings for both the throttle and brake... they were the same as before)
- I double checked again the wire colouring and orientation of the -ve and +ve probes. Same result as before.
- For the throttle, I put the limit/stop at the maximum on the pedal, and took a new reading when fully depressed.
- For the brake, I took a new reading again and really leaned on it. The Smart Control app indicated a max pressure I could put of about 100kg. I will NEVER be able to put that much pressure when actually using it in game. No way. I change the description to 100% since that's really all I could manage. (I'm surely not as strong as others, but this is life)

 

[GeekyDeaks]

- I double checked again the wire colouring and orientation of the -ve and +ve probes. Same result as before.

It is very odd that HE would have this completely backward, but I can see from your readings that it's consistent across both the supply and the signal. It doesn't matter too much though as all you are really interested in is the voltage difference across S-/S+ and you'll be doing your own wiring.

For the brake, I took a new reading again and really leaned on it. The Smart Control app indicated a max pressure I could put of about 100kg. I will NEVER be able to put that much pressure when actually using it in game. No way. I change the description to 100% since that's really all I could manage. (I'm surely not as strong as others, but this is life)

On the positive side, this shows the range on the HE brake pedal is pretty close to the CSPv3, however the fact that it requires so much force means we need to find a way to get the same mV for a lower force. One way might be to increase the mechanical advantage on the pedal so that it hits 100% easier, then recalibrate on the PC side (as you can easily do that). Does the ulitmate allow you to adjust the arm to do this?

I found a document with some hints about bridge balancing that might help too:

https://www.technimeasure.co.uk/wp-content/uploads/2016/03/Introduction-to-Strain-Gauges-and-their-usage.pdf

 

[Atak_Kat]

the fact that it requires so much force means we need to find a way to get the same mV for a lower force. One way might be to increase the mechanical advantage on the pedal

The way the HE are designed, I don't see a way to do that, and honestly I'm not so keen to go down that path. 80% of the time, I'm sure I will use the HE with PC, and I don't really want to mess with that. The idea here is to find a solution to enable the HE feeding the CSPv3 board, for use mainly when I connect my wheel/setup to PS5 (not often)

For the range, I have a hope it will be fine as-is. I had the CSPv3 pedals setup basically as heavy as they would go, much heavier than they would come out-of-box. So 'easy to max out' on the CSPv3 is relative to the HE. But I suspect many users have them setup with far less pressure range. So I have a feeling that the CSPv3 board will be able to adapt to ranges lower than what I tested. And since this is only really for using with the PS5/GT7, I seem to remember that GT7 auto-calibrates anyway. So, it's worth a shot as-is (for the brake)

Breadboard, RJ12 cables (6P6C) and the little load-cell amps arrived today, so I'm going to venture down the path to make an adapter cable between the HE brake output, and the CSPv3 brake input. I don't have a crimping tool, so it's gonna be old-school wire splicing, soldering, and heat shrink. Gonna be a bit ugly, but if it works, we'll figure out how to make it nicer at a later stage.

Regarding the amplifier suggestions, I couldn't really find the ones you wrote in stock at local on-line suppliers. But there are thousands of options. If you can give me some hints how to filter for important 'specs' I can certainly check what's available.
- My assume filters are, in-stock, rail-to-rail (don't really know what that means though), through-hole mounting, number of circuits = 1. There so many other filters I really don't understand. But if I know what to look for, I can list a few options easy to get here.
- For whatever reason, this one seems to often be near the top of any list. NO IDEA if that's a good option or not.

 

[GeekyDeaks]

The way the HE are designed, I don't see a way to do that, and honestly I'm not so keen to go down that path. 80% of the time, I'm sure I will use the HE with PC, and I don't really want to mess with that. The idea here is to find a solution to enable the HE feeding the CSPv3 board, for use mainly when I connect my wheel/setup to PS5 (not often)

Yeah totally get it, just throwing ideas out there

rail-to-rail (don't really know what that means though)

It means that the amplifiers range is the same as the voltages used to power it. Since this is just DC stuff we are messing around with, then for our purposes, when the CSPv3 is supplying 3.27v to the amp, it means we can actually amplify the output all the way to 3.27v rather than say 2.9v...

I'm off out today, but I'll take a look at the datasheet for the AD623ANZ when I get back to see if it's suitable

 

[Atak_Kat]

THE BRAKE WORKS!!!!


Some notes related:
- As we suspected, the signal seen by the Fanatec Control panel is lower than would be with just the CSPv3 brake connected (for similar force). I had the Brake Force set at 100% with the CSPv3 pedals input, and as we suspected, with the HE connected that would require tremendous brake pressure to get to full 100%. But with the HE brake connected, I reduce the Brake Force (in the Fanatec Conrol Panel app) to about 50% and this gives a full range of HE brake graph, with 'reasonable' but heavy force applied to the pedal (something more realistic while really in-game)
- I made the cable/connections (ie. E+, E-, O+, O-) inline with what my testing showed (not the info given by HE). The pedal operated properly as far as I can see.
- For future reference below are the connections used for the cable made.

LETS BE CLEAR for anyone following and maybe trying this on their own:
- Although I'm careful, I can deal with the reality that I might mess up. I don't blame anyone for my errors (neither judgment/trust, nor execution). If you mess up trying this, that's on you. It's possible my diagrams have errors, so check yourself and this is not a how-to guide.
- It is strange that the HE wiring colours and related +/- seem to be opposite of what they informed me. That might be for many many reasons and how they design their own circuits (or, just my error from the beginning). Maybe the pedals I have are older and an anomaly. Regardless, the info below should be according to what I tested/learned, and surely with the background of a pure novice that will make many errors.

READY to tackle the throttle and clutch next-steps with the absolutely great support of GeekyDeaks or anyone who may have further guidance/suggestions.

 

[Atak_Kat]

Interestingly, the hall sensor is also a wheatstone bridge, so that also has the MCP6N11 amp, the only problem is that I think that amp only comes in surface mount packaging.

I am dangerously curious about this point... I wonder if there is some way to sacrifice/use the components on the hall-sensor board towards the target????

 

[GeekyDeaks]

But with the HE brake connected, I reduce the Brake Force (in the Fanatec Conrol Panel app) to about 50% and this gives a full range of HE brake graph, with 'reasonable' but heavy force applied to the pedal (something more realistic while really in-game)

That is brilliant news about the brake pedal and that you can adjust the range. Does this also carry over to when you use it on the PS5? I'm curious if it's a setting on the microprocessor on the board or on the driver software on the PC. If it saves the settings on the microprocessor then that is one thing that doesn't need to be addressed in hardware and if you can do similar things on the accelerator and clutch it should also help simplify things there too.

It is strange that the HE wiring colours and related +/- seem to be opposite of what they informed me. That might be for many many reasons and how they design their own circuits (or, just my error from the beginning)

Funny you should say this. Whilst I was out I was thinking about both this and the fact that the red wire is GND on the Fanatec. I honestly don't think you are doing anything wrong as it would have produced odd results when testing on the hall sensor board, but if it was me, for piece of mind I'd take a battery and measure it to be sure it reports a +ve voltage as expected.

For whatever reason, this one seems to often be near the top of any list. NO IDEA if that's a good option or not.

That amp has a maximum gain of 1000. If the adjustment you tried on the Brake also works on the clutch and throttle, and still works on the PS5, then it should suffice. If not, then you are probably going to require something with a larger gain as you have a 2mV change on the throttle that needs to be amplified to 3.3v which is a gain of about 1666