Electronifying A Horror Fraught Hydraulic Press

[Josh] is replacing the springs in his car’s suspension. He wanted to know the travel rates of these springs, but apparently, this is a closely guarded trade secret in the industry. One company did manage to publish the spring rates, but they weren’t believable. Instead of taking this company’s word, [Josh] built a spring tester.

The theory behind a spring tester is pretty simple: apply a force to a spring, measure it, then measure how much the spring has traveled. Or compress a spring an inch or so, measure the force, and compress it some more. Either gets you the same data.

This spring tester is built around a Harbor Freight hydraulic press. Yes, the spring is completely captured and won’t fly out of the jig if you look at it wrong. The bottom of the press contains a few load cells, fed into an ATmega8, which displays a value on an LCD. For the displacement measurement, a ruler taped to the side of the press will suffice, but [Josh] used a Mitutoyo linear scale.

What were the results of these tests? You shouldn’t buy coils from Bilstein if these results are correct. The rates for these springs were off by 70%. Other springs fared better and won’t bind when going over bigger bumps. That’s great work, and an excellent application of Horror Fraught gear.

38 thoughts on “Electronifying A Horror Fraught Hydraulic Press

    1. I do wonder if the 70% figure he arrived at, is within the margin of error. Is the load cells response linear? What is their calibration? But I do wonder why mfgr’s won’t supply relevant data for their products.

        1. TBH aftermarket most don’t understand or care about rates, they just want a sick lowering kit, bro, or “better handling”, or just shiny coloured springs.

          Some manufacturers do publish actual data, Bilstein definitely do for some of their shocks as I bought a set of a specific spec. Land Rover do (or did) publish all their spring rates in their service manuals.

          Bilstein make a whole range of things at a whole range of prices, and they’re not exactly a bunch of cowboys. I’d suspect a home-made load cell over a vast OEM spring manufacturer…

      1. depends on the load cell i suppose. the ones I use output 0-30mV and are linear (would it make sense to have any other response?) I have gotten them calibrated but they were still well within spec after many years of use. I would actually be more concerned with the output of the amplifiers.

  1. An awesome build! Well executed methodology.
    However, when I had my coilovers rebuild by King Suspension, not only did they replace the springs; they phoned me to tell me they had run the old ones in their dyno and found they were non-normal rate. I was impressed they bothered to check the springs at all, let alone phone me about it.

  2. I’m assuming something is wrong with his setup – I’d be willing to be a company like Bilstein isn’t going to be 70% off on their spring rates. Also since when are spring rates a closely guarded secret? Whenever I’m looking for motorcycle , snowmobile, etc springs the rates are always clearly defined so ‘you can pick the rates you need for your application.

      1. *is joke.
        There is quite the possibility that the homemade test rig is inaccurate. Best to test several different brands of springs (if possible) and plot the data, à la Mythbusters.

    1. interesting thoughts there… let’s assume that Bilstein is making these springs completely according the specs (these things happen). Or perhaps some faulty spring (or the wrong one) has been used to do the measurements. And some wiseguy does some complicated measurements but somehow manages to make a mistake (these things can happen). However the results are published… on the internet, via the very (in)famous website hackaday, people read this and immediately believe it and spread the news… wildly!!!
      I do not think that Bilstein will be amused as it is their reputation that is at stake here…
      Because hackaday is supporting the spread of this news, could this (via huge lawsuits) lead to a situation where the hackaday website would be forced to close/shut down?

      What I mean is that you can shout/promote anything on this website about diy projects. For instance: how to laser your hands, create electrical arcs of many inches long, spring loaded knifes or war machines. Or anything else that’s dangerous to your heath, but you just can’t shout that some products of some firm are crappy… (first saying that specs are not/badly available, then saying that a product doesn’t meet specs) you just don’t do that!
      I’m afraid that this will lead to an army of lawyers that will hunt you (for the article) and me (for reading and commenting) will find us… (or hunt us down if they need to) and sue us for everything we’ve got (or should have).

      1. This isn’t how science works. You don’t *not* speak out about the results of an experiment for fear of retaliation.

        Someone studied and published results. And now people (peers) can review his work, point out where flaws might be, or choose to do an experiment of their own and see if the results are repeatable. Maybe those who come after will find a more accurate way to measure the distance or the force in a meaningful way.

        If the methodology or result was wrong, Bilstein can also join in with the science. Instead of going to court, they could spend less time and money by showing why they’re the leaders and why their products are to be trusted. Their biggest win would come from them performing a similar experiment of their own, showing their results, and suggesting how others can get the same results. Now *their* technique can be reviewed, too, and science moves forward.

        On the other hand if it’s correct and that particular spring didn’t measure to the stated specs, Bilstein could still join in and give credit and help find an explanation for what happened.

        1. Sadly, that is how science is done in many fields. Try publishing something about global warming, or any number of other controversial, politically correct, or industry-guarded areas – there’s only one correct view on it, and your data better not be an outlier that suggests otherwise.

          1. Well.. no. Not really.
            Your description is more in line with how POLITICIANS and the uninformed public often views science: that their interpretation/view is right and everyone else is wrong. Through a combination of a lack of understanding​ of how science REALLY IS done, and the egocentric desire to be ‘the man with the plan’

            The only real concern which can restrict science being peer reviewed, confirmed or busted is when it is entirely controlled by a business or industry with a vested interest in having their way. “Study shows no link between cigarettes and lung cancer!” (Study funded entirely by a Tabacco conglomerate, and consisting entirely of a survey handed out to smokers “do you now or have you ever had lung cancer?”)

    2. “I’d be willing to be a company like Bilstein isn’t going to be 70% off on their spring rates.”

      They’re progressive springs, so it’s entirely possible that Bilstein is quoting the spring rate at a different load or an average over a range (or possibly that they recommend a different ride height, I dunno). He’s interpreting it as the rate at normal load (ride height), but they might be taking a larger range. The springs basically transition from ~180 lbs/in to ~366 lbs/in, so depending on where you defined it, it could be correct. (Not that that makes them good, mind you).

  3. In the pic shown, the first few coils are already bound together, which leads me to believe that this is a progressive rate spring, not a linear rate spring. If it was linear rate, all of the coils would come together at the same time.

    In a progressive rate spring, there are two different spring rates to consider and the higher spring rate doesn’t really come into play until the coils of the lower rate section come together, which would happen during periods of higher suspension travel.

    The reading that [Josh] is seeing in the image above is likely the lower of the two rates. Compressing the spring into the higher rate section will yield the higher rate. Some companies will give both rates, some will only give the higher rate. Generally speaking, you’ll only see progressive rate springs in OE applications that fit the stock style struts. Coilovers, on the other hand, almost always use linear rate springs that are labeled with the actual rate.

  4. As far as I understand, Bilstein, as a manufacturer of suspension products is held in very high regard. They are my choice for the next set of shock absorbers for my FJ40. I would have great respect for them if I saw them rebut and explain the discrepancies shown by this experiment, if this test was not fair or accurate I would love to see them take the time to show us why. At the same time, I would be very disappointed to learn if they were to sic the lawyers on him or websites that are covering his experiment, to the point where I would choose a different supplier for my rig.

  5. I used to work in a Metrology Lab (for the uninitiated, we mostly calibrate measurement instruments) and we routinely serviced and calibrated several testers that worked exactly like this build.

    Key difference though-force was applied via a precision jack-screw/servo motor assembly so expansion-type springs could also be tested, and the actual criteria was the measured distance of compression/expansion vs. force measured at the head (it was reversible for compression/tension measurements).

  6. Looking at his chart it you can calculate that the spring rate is actually ~192lbs/in for the lower rate and increases to ~326lbs/in for the higher rate. 192 is only 10% off the advertised spring rate of 215. If his car’s ride height compresses to the spring’s higher rate immediately, then either he bought the wrong springs or he has a defective spring. Conveniently thou he tested the springs to show they are only 10% off. I suspect he has the wrong springs for his application.

    1. The bilstein shocks have 2 spring rates depending on how compressed they are. You can visualize spring rate by looking at the slope of this data. From 0-2.25in the rate is ~192lbs/in and from 2.5-4in it is ~326lbs/in. Where did you get 245 from?

  7. Oh, maybe he’s mixing up terminology. The compressed height at a set force is 70% different from stock? (One of )The point with coilovers is you can adjust spring pre-load w/o changing ride height?

    1. Yeah, I dont think he understands working rate. He states “I’ve gathered that “working rate” means rate at ride height”. If your car’s ride height is 3.5in that does not mean the springs have compressed 3.5in only that the car’s geometry places the car 3.5in above the ground. Instead he should be looking at his car’s corner weights which will tell him the static load each spring is supporting.

      Also adjusting spring pre-load will change the ride height without changing the spring rate.

      1. Yeah, but even doing a *very* low-ball guesstimate on that puts it pretty close to what he was saying. Unless I’m missing something? Assume a car weight of 3000 lbs (underestimate) evenly distributed on 4 wheels (way underestimate), so each spring is carrying 750 pounds. At 180 lbs/in transitioning to 366 lbs/in after 2.25 inches (underestimate again, it’s not a sharp transition), to carry 750 pounds you’d compress 2.25 in to carry 405 lbs, and then an additional ~1 inch to carry the remaining, puts you at 3.25 in, where the spring rate’s 366 lbs/in.

        1. If the car weighs 3000 lbs then yes this is how he calculated the spring rate.

          He has to be using the wrong springs. The intent of progressive springs is to be in the lower spring rate region under the static weight of the car and enter the higher spring rate region when driving over bumps.

  8. I used to work with MTS servo hydraulic systems and did scientific instrumentation. Even though the setup looks simple there are a lot of places where things can go wrong, and if your numbers are off by 70% something is very wrong.

    If I were in your shoes, the very first people I would contact Bilstein and ask them how they do their test, and show them how you did yours. I put more faith in their numbers than yours. I don’t know anything about the standards for testing shock springs but I suspect there is a standard and I suspect they adhere to it.

  9. Always check the setup with dead load. There is no great difficulty in piling up sacks of material of known weight to make up a decent test point with a known load.

    Also, coil springs have been analyzed to death. They are operated by torsion of the wound coil and by the moment arm of the diameter of the spring. For a given material the shear elasticity is well understood and the section properties are easy to determine. It’s the last characteristic that gives the most variability as (it’s been a while and I’m not looking it up) it is relative to the 4th power of the diameter (could be cube, but I don’t think so) which means it is very sensitive to wire diameter. Which means the makers control the wire diameter. The travel limits are most affected by the pitch of the spring, which also limit the maximum shear load. Put too few coils in and the spring will fail in torsion before it bottoms out.

    In the progressive spring they allow a small pitch at one end and a larger pitch at the other so that more coils initially are in series (like capacitance) making the force/displacement lower. Then the initial gaps close removing coils from the displacement path (already displaced as possible) effectively making the force/displacement higher.

    If you want ridiculously variable springs, look at belleville washers. Those are very sensitive to thickness, distributed internal stress from forming and amount of coning. Like +50%/-70% variations from catalog values.

    But coil springs? +/-10% would be a surprise for anything that carries a rating.

    Fun fact – spring rate is not affected by hardness/strength. Hardness/strength only affect how far the spring deflects before it plastically deforms; so if you need a lot of travel it had better be strong.

    The Bilstein is listed as a suspension lowering kit. I don’t see how he’s surprised it’s solid height is reached before it gets to the factory bump stop. The experimenter picked the wrong spring for the application.

    Finally, there are plenty of spring makers in the world. They have spring testers and would probably charge not very much to do a better job, plus they should have guards in place in case the spring ruptures from some internal defect so no one will be stabbed with high-speed shrapnel.

    tl;dr Experimenter shows the working range is 190 lbs/in on a spec’d 215 lbs/in, then complains that variable rate springs have a variable rate. Then complains that overloaded springs push him into the second stage of the variable rate. Smart guy with wiring; not so good with mechanical engineering. Also experimenter should have asked Bilstein for the load/displacement graph to see if it matches what he made before publishing his 70% diatribe.

    1. Check out the deflection on the aluminum base plate under the spring…not a very stable or rigid setup. Looks like he is measuring the springiness of more than his car springs.

    2. +1.

      -He also makes no mention of suspension geometry which can significantly affect the force applied to the spring.

      -He states I Quote “Note that I’m not doing the typical 2″displacement minus 1″ displacement equals rate.” If he is not doing this then what is he doing? Averaging the rate from full displacement? This shows a lack of understanding of how progressive springs are intended to work.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s