Eddy vs. hysteresis brake theory 1

[amorrison]

Can anyone provide Comparison information on eddy brakes and hysteresis braking operation.

From Googling - they are quite different as eddy current brake forces are ~ proportional to rpm while hysteresis brakes provide forces proportional to the brake electrical coil current.
An "applied physics" answer would be nice.
Thanks.

 

[rmw]

I googled both, and did some reading, and from what I can ascertain, the eddy current brakes used a fixed current in a single or parallel sets of coils, which sets up a fixed magnetic field(s) through which the rotor passes, where the torque is strictly a function of the speed of the rotor through this fixed field, while the hysterisis brake varies the current flow through the coil via a controller to produce a controlled torque in the rotor.

As I see it, one could take an eddy current brake and disconnect it from its constant voltage source (producing the fixed current and hence fixed field) and connect it to a controller getting its process variable from a torque load cell, producing as its output current proportional to the PV, to control the torque, and the same device would then be a hysteresis brake.

Now, I could be wrong, but that is what I got from the reading.

One thing I read from the first listing on Google for -eddy current brake - was some student postulating that such a thing was used on train wheels. I hope not. The eddy current brakes that I use get the rotors cherry red hot, and it seems to me that a train wheel operating cherry red hot wouldn't have the same strength rating as a cold wheel.

Our eddy current brakes have the rotors designed such that as they reach maximum temperatures, they become cherry red. The rotor is designed such that it as it grows thermally it expands away from the magnet pole plates, opening the air gap, and reducing the induced eddy currents. The coil current is also affected over time by the I^2*R losses which increase the electrical resistance of the coil wiring, and if not compensated for by a corresponding increase in the voltage, reduces the strength of the field. Usually by then the rotor is cherry red hot anyway, so it doesn't want any more heat.

I have seen air gaps set at 1/8" cold open up to 3/8" hot in a steel rotor that wouldn't blink if a train ran over it.

If someone can explain it to me in "applied physics" then I could be counted among the learners, too.

rmw

 

[Warpspeed]

I too have been mystified by the theory behind eddy current dynamometes, which is why I find this topic so very interesting.

Originally I believed that the electromagnetic coils just set up a fixed magnetic field, and the motion of the rotor through the airgap induced eddy currents in the rotors which increased directly with rotor Rpm. This idea came from the observation that generators and alternators always increase their electrical output with increasing speed with fixed field excitation.

This appears not to happen in a dynamometer head, and braking torque remains fairly constant over a fairly wide operating Rpm with constant ampere turns of excitation. When trying to reverse engineer commercial eddy current dynamometer heads, the calculated flux density in the magnetics generated from the specified ampere turns, always works out to be way beyond the saturation flux density in the steel in the magnetic path.

What seems to happen is that the excitation coils set up a magnetic flux through the whole machine, and eddy currents circulating in the rotors set up an opposing flux, the resultant magnetic flux being far less than it would be with stationary rotors. The more ampere turns in the coils, the more torque that can be held, seemingly without saturating the steel core.

A mechanical analogy might be a fairly flimsy mechanical diaphragm with 1,000psi of air pressure on one side, and 1,000 psi of oil pressure on the other. The diaphragm itself would hardly be stressed.

Likewise huge total ampere turns in the excitation coils, and massive circulating rotor currents do not seem to lead to magnetic saturation in the machine.

Rmw is correct in saying that the typical eddy current brakes used in trucks are designed to be self limiting, the rotors dish outwards with heat, increasing the air gap, hopefully preventing a total meltdown at sustained high load. Heat buildup in the excitation coils will also reduce the excitation current (at a fixed applied voltage) by increasing the electrical resistance.

Another observation is that the excitation coils have very high inductance and low electrical resistance, T=L/R so the time constant will always be fairly long, perhaps half a second or more. That will make it rather slow to react which is a definite advantage in a vehicle, but perhaps not so good for a workshop dynamometer.

 

[rmw]

Warpspeed,

If the dynamometer head you refer to is the dynamometer version of the truck/bus retarder, then I don't understand your third paragraph, unless you are referring to the condition where the flywheels are heat soaked, and the coils are at high temperature, and the voltage source, normally 12/24 V. batteries are depleted.

The torque curve of the truck/bus retarder is much like an engine torque curve with a peak at about 25% speed range, and a constantly increasing horsepower curve. This is in the cold condition.

Both of these curves flatten somewhat in the hot condition. I have witnessed them being performance tested where they are made in Europe, principally France and Spain.

The flywheels for this type of equipment survive many heat/cool cycles before they eventually heat check, crack and fail.

Another key to that device's success is the metallurgical selections in the rotors and stators that prevent residual magnetism which would cause parasitic drag on the vehicle when the retarder is not in use.

If you are using the truck/bus style retarder for a dynamometer head, you can turn the flywheels around and rotate them backwards. They are designed as backward inclined fans so that they are non overloading at higher speeds to reduce parasitic loads at higher vehicle speeds.

Turned around, they approach forward curve sirroco type fan designs, in which the horsepower curve is constantly increasing, but which, in a dynamometer application is beneficial. The cooling is much better as a forward curve fan design.

Actually, reflecting on your second paragraph, what you state in your first sentence is exactly what is happening in the case of the truck/bus application in a vehicle or a dynamometer. The coils are wired in pairs so that every other coil pole shoe is north pole, while the odd numbered coils are south pole. So, depending upon how many coil sets there are, 6,12,8,16 or two, (the two coil models have pole shoes as part of the stator that set up 5 N/S pole pairs on each flywheel) there are that many north south current fluctuations in the flywheel current as it rotates through all those fields. Hence the eddy currents and the drag. Imagine what your generator/alternator would do if you shorted the output leads.

They work wonderfully, but alas, in the truck/bus industry they have what is dreaded the most, they are heavy.

rmw

 

[Warpspeed]

I am most familiar with Telma units, most commonly fitted to Scania busses and trucks. As you say, the torque curve follows something that looks pretty much like the torque curve of some sort of typical engine. It is not flat, but "reasonably" flat.

The torque most definitely does not just keep rising with Rpm in some sort of exponential curve as the output from a generator or alternator would. I am not sure exactly why, and that is all part of the mystery.

 

[rmw]

I think part of it is as has already been discussed. I was once involved with a version of the same concept of retarder that used a Telma type, (they are the king of the mountain, and specialize in the 16 coil type with the backward inclined bladed flywheels) or the two coil type with some more radial bladed type flywheels made by a variety of mfgr's in Spain, that was and is used in semi-trailer axles, and users of the product reported that other truckers would often get on the CB radio and tell them they could see something burning under their trailer.

They were seeing the flywheels glowing red hot. I have also installed several of the drive line versions of the two coil style in truck tractors, and in order to prove that any driveline vibrations that subsequently appeared weren't the fault of the retarder, as well as to prove the retarders sheer power to skeptical owners, I would complete the installation, all but the final drive line to the rear end, and would then run a 'chassis dynamometer' demo.

I would instruct the operator to put the vehicle in a top gear in order to get the driveline speed up, and then start stepping into the retarder one step (out of 4) at a time.

The drag on the engine was rather instantaneous, and the appearance of the reddening of the flywheel faces, along with the widening of the air gap. Normally, the truck engine wouldn't pull it in the highest gear. A L-1, L-2 gear would have to be selected in order to pull the retarder at max engine speed at all.

But, that said, the actual dynamometer testing of the retarders to develop the cold (and hot) torque curves is done under controlled procedures where the effects of the heat fade are minimized by letting the retarder cool between data points, and by taking data at various speeds with the retarder cool at each data collection point.

So, someone smarter than me is going to have to tell us why the torque curve is curved as it is with its peak well below maximum speed, and tailing off fairly rapidly at maximum speed. One great thing about that characteristic is that just like a high torque rise engine, the retarder 'pulls' at lower speeds, and will provide good hold back at lower speeds, while absorbing enormous amounts of braking horsepower at higher speeds. The horsepower curve is fairly linear across the speed range.

I think the answer might be in my challenge above. What would the torque curve of a generator/alternator look like with a shorted output where the power generated never left the internals, and they heated up rapidly.

BTW, the two coil versions of the same gadget seriously out performed the 16 coil versions, and could be had in copper coil wiring instead of aluminum. Much better for something running under a vehicle in our country where lots of road salt is used in the winter time. And, they drew about 3/4 of the current (180 amps vs 240 amps) of the 16 coil version in 12V applications to produce that outstanding performance.

rmw

 

[Warpspeed]

Great stuff rmw, I am not familiar with the two coil type of retarder. Who makes these or where should I begin looking for further information ?

I have been giving some serious thought to this rather odd shaped torque characteristic and have a hypothesis, only a guess really, but it may possibly explain the odd torque/speed curve.

The coils are energised with direct current, and so the magnetic MMF created will be constant. Any change of characteristic with speed must therefore occur within the rotors themselves.

At zero speed there will obviously be zero torque holding capacity, there must be some relative motion to induce eddy currents in the rotors. These currents will generate their own magnetic field that opposes (repels) the field generated by the field windings. This repulsion is what makes it harder to turn, and the resulting eddy currents produce heat in the rotors.

With an increase of speed above zero, the torque required to turn the retarder will rise rapidly to a peak, typically at about 800 Rpm, at least for the retarders I am most familiar with. Then something happens at higher speeds to prevent this torque rise from building further at ever higher Rpm.

In transformers, motors and generators, where there is a very fast change in current (Amps/second) there is a phenomena called skin effect. The current flowing creates its own varying magnetic field which forces the current to flow in the outer skin of the conductor. This is highly frequency dependant and has the effect of raising the electrical resistance of the conductor.

http://www.allaboutcircuits.com/vol_2/chpt_3/6.html

or Google "high frequency skin effect" for more info.

What I think might be happening is that the circulating eddy currents in the rotors may be forced to the surface of the discs as the Rpm increases. This would effectively increase the electrical resistance of the rotors, and reduce the braking effect as speed increases. Only a theory, but I cannot see any other mechanism that would account for this torque fall off at high Rpm.

The skin effect is a real frequency related dimension, and the way to overcome it is to use relatively thin conductors suitable for the operating frequency. In an electrical generator or alternator the current flows in wires that are relatively thin with many turns, so skin effect is not going to be a problem except at extremely high frequencies or Rpm. So shorting out a generator or alternator would indeed create an ever increasing load with increasing speed, probably up to the point where the windings burned out.

But an eddy current retarder has two very massively thick discs, and if the current only flows across the surface to a small depth of (say) a few mm, that would definitely increase the effective series electrical resistance significantly as speed increased.

These seemingly simple eddy current retarders are far from simple !!

 

[rmw]

I am pondering your hypothesis. My immediate observation would be that the flywheels I have seen, and that is many, are heat marked completely through the discs, as well as cracked all the way through on those that have seen severe service. The heat marks-the tan color steel gets when it has been heated to cherry red and cooled-go all the way through and slightly up onto the fins or fan blades.

Is there a maximum current density type of effect (I am a mechanical engineer, excuse the terminoligy) where up to the 800 rpm speed the eddy current flow is not limited, but afterward, it is??????

Since the eddy current changes direction 4 times each revolution, isn't that an alternating current type situation, and as the flywheel rotates, wouldn't there end up being a lag to how the current flowed in regards to the pole passes, and the faster the rotation, the farther back the lag????

I am grasping, but this discussion has peaked my curiosity. I never pondered why the torque curve had a peak, and not a steadily increasing value. I am really curious now.

Here is a link that shows a 2 coil type retarder mounted in a trailer axle.

http://www.tmm.es/webtmmi/default.htm

Look at the iron in that thing. Copper coils, too, 2 of them. They weigh 40 lb. each.

If you are only familiar with Telma, since they use only one style flywheel, take a look at this variety.

http://www.tmm.es/webtmmi/default.htm

I looked all over the internet for 2 coil retarders. I didn't find any. I know that they were made by Frenelsa and Cofresa, because I have bought units from each of them. I suspect that they don't show them as product listings because they are so heavy. That is the reason for the 16 coil retarder. Light and cheap. I don't like them. Aluminum wire, with all the problems associated with aluminum wire, and lots of connections to corrode and loosten.

I know the axle manufacturer manufactures them for their axle product, but I don't know if they make it in a driveline version. I think they could if asked. They do have the designs.

rmw

 

[Warpspeed]

I would expect the heating and cracking to be fairly uniform, because even if the heat was mostly generated right at the surface, it is also the only place where the heat can escape from. The bulk of the disc will probably just get fairly uniformly hot through thermal conduction.

With a sixteen coil Telma there are four alternate magnetic circuits, just as you say, and at 800 Rpm that will be 3,200 cycles per minute, or a cyclic frequency of 53Hz.

Digging deeply into my electronics reference books I find a formula to calculate skin depth (in copper conductors only unfortunately) that comes up with a figure of around 9.6mm penetration depth at 53Hz. Now maybe it is just a coincidence that the thickness of typical rotors might be very roughly about twice that at the peak of the torque curve.

It suggests to me that at Rpm <800 perhaps the current distribution is fairly even throughout the bulk of the disc material. At >800 Rpm, the current may only then begin to crowd towards the outer surface at an ever increasing rate.

If all this is actually true,(and maybe it isn't) thicker rotors would create their torque peak at a lower Rpm, and thinner rotors at a correspondingly higher speed.

One further observation that appears to contradict all of the above is that as you switch circuits in and out of a Telma, the torque peak steps higher or lower, but still occurs at the same Rpm. That appears to contradict the above theory, but maybe not.

The concept of frequency here is instantaneous rate of change rather than cyclic repetition rate. In other words if the retarder only has one circuit switched in, it still sees a rate of change equivalent to 53 Hz, but then rests for a while before seeing another equivalent fast rate of change.

I really am intrigued by all this as well, and would dearly love to know why it does what it does.

 

[rmw]

The interesting aspect of what you have proposed is that with the Telma type, as we will refer generically to all 16 coil models, since they invented that style, as circuits are incrementally switched on, the N/S pole combinations also increase incrementally, resulting in the individual performance curves as you correctly describe them.

In other words, with only one circuit engaged, there is one north pole, and 180 degrees away, a south pole. When the second circuit is engaged, now there are two north poles and two south poles, each 180 degrees from each other, and 90 degrees from each other. Ditto with circuit 3 & 4.

With the two coil version, since the pole shoes are built into the iron of the stator casting, the Spanish word for the piece that bolts on and holds the coil in its proper place is 'star', there are 5 pole pair, since the 'star' has 5 pole shoes cast into it, and the central casting has 5 pole pieces cast into it as well.

In the link I gave for the retarder axle, you can see the pole shoes of the star right about on the horizontal centerline of the retarder, adjacent to the electrical connections. The iron masses that extend to the flywheels above and below that are the pole shoes of the central casting portion of the retarder core.

Therefore, the flywheel always sees 5 N/S poles, only the flux density is different depending upon the number of circuits energized it the two coils. The coils of the two coil retarder each consist of 4 parallel simultaneously wound wraps of coil wire. I used to know how many turns there were, but can't do it from memory anymore. I had done some ampere-turn work years ago, but can't remember the turn count.

But, other than having an overall higher torque rating, and maybe now I know why, the torque curves of this style retarder are the same as the 16 coil (or 8 coils or 6 coil) models.

I will tell you this. Once, when in a pinch, flywheels from the 16 coil type retarders were used in place of those intended for the 2 coil version (the 16 coil retarders for the axle retarder have 270mm flywheels while those for the 2 coil models are 300 and/or 310mm. The two coil retarder trashed those flywheels in short order. I always attributed it to the reduced amount of mass of the iron discs of the smaller flywheels turning in the flux field intended for the larger flywheels. They were telma type, by the way, and that is how the retarder mfg'r refers to them. The 310mm flywheels are called the iruña type for the company who originally designed them.

If you look at the TMM link that shows the selection of retarders you can see that type flywheel. It is the lower center. We also call it the 40 fin type for English only speakers. You can see the additional fins and the closer to radial nature of their layout on the disc of the flywheel.

A telma type flywheel for the retarder axle will weigh right at 100 lb. while the iruña weighs in at 125 lb. 25% more mass in 30mm difference in dia.

In light of this, where does that put the hypotheses? I find this to be a very interesting discussion, and hope others will jump in and shed some light on this for us.

rmw

 

[Warpspeed]

I too am a bit disappointed nobody else has joined in this discussion, even amorrison4 seems to have gone to ground. But this is a very highly specialised field and I am beginning to think that not very many people truly understand how these eddy current retarders actually work.

My background is mainly in electronics, and although I have done a very little magnetics design, magnetics is rather a black art even for most hard core electronic people.

A typical Telma coil (one of sixteen) consists of 380 turns of 2mm aluminium wire and may have around 17 amps flowing through it when connected across a 12v supply. There will be four coils and four airgaps in each individual magnetic circuit, all in series. The ampere turns per coil is going to be about 6,460, and that would be sufficient to create a magnetic field strength that should saturate the steel core many times over.

The other puzzle is the shape of the actual torque absorption curve with varying Rpm, and fixed dc excitation current. From your comments I take it that the two coil version has essentially very similar speed/load characteristics to the Telma type. To me, that hints rather strongly that it must be rotor design (thickness?) that causes this characteristic torque shape.

What country are you in rmw ? I am from Melbourne Australia.

 

[rmw]

I am in the USA.

I think those that know the answers aren't on this fourm and those that I know don't speak english. Telma might have some english speakers that do, but the Spaniards that I know that are the older technical gurus don't, if they are still alive at all.

Later in the day I will get the ampere turns for a typical two coil style coil. I have drawings, but they are way off my beaten path. Maybe I can find them.

One generality I learned as a ME early on when taking those hated EE courses, (not nearly as much as they hated taking thermo) and possibly some early generator training was that the more iron and copper it had in it the stronger it was with respect to magnetic field strength. After all, that is what does the work in the flywheels, irrespective of their design.

I was also disappointed that I couldn't find any more references to the 2 coil style of retarder on the web, but with the advent of the borderless EU, now that gross vehicle weight matters to everyone, those things, once real popular in Spanish intra-country transportation where Spanish laws favored their use, and negated their weight penalty, has given way to the 'god' of light weight, even at the sacrifice of performance. That thing may have peaked and disappeared before the advent of the modern web.

rmw

 

[Warpspeed]

I would not be at all surprised if the guys that did all the original research and design are all now long gone. Searching out the original patent papers may be a fruitful line of inquiry. There are features and ideas such as the deliberate dishing outward of the rotors with thermal expansion that can only have been developed empirically through practical testing.

As with any kinetic brake, size is important for thermal considerations if nothing else. As you say, the magnetic and physical loading of a larger unit is going to be lower than a smaller device. An unfortunate consequence of all that mass of iron and copper is the stored magnetic flux, and the very slow response speed to electrical input.

I have in the past attempted internet searches about these eddy current absorbers, and detailed technical information is extremely difficult to come by.

Here in Australia most long distance trucks use Jacobs engine compression brakes rather than the eddy current retarders, so they are fairly uncommon here. The biggest use here in oZ is probably for chassis dynamometes. That field is dominated by one single company, and they are renowned for being very secretive and uncooperative. Not much help available there unfortunately.

 

[rmw]

Telma invented the eddy current retarder shortly after WWII and has always appeared to me to be the "chevy" of the retarders. They offer a one size fits all style that optimizes mediocrity.

The 16 coil aluminum wired retarder with all its connections is a maintenance nightmare if run on roads where lots of salt is used. I now see other manufacturers offering copper wire coils.

Aluminum wire has its own bad characteristics. It corrodes, it cold flows, meaning that terminations losten up over time. Many a house fire in our country is caused by aluminum wire connections lossening up and getting hot.

Their backwards inclined flywheels will always be the least horsepower hogs at higher speeds, but they don't cool as well as others. As a dynomometer I would always run their flywheels backwards for better cooling and additional parasitic horsepower.

If mimickery is the best form of flattery, then they should feel well flattered. Their unit has been cloned many times by other companies. There are manufacturers whose units are virtually undistinguishable from Telmas.

The flywheels are all cast in the same few foundries. I once took a Telma dealer from our country (now former Telma dealer) to a foundry in Spain to see where another manufacturers flywheels were made, and he was much surprised to see the foundry full of Telma flywheels for Telma models that he well recognized. There are no secrets in the metalurgies used by any of the manufacturers. Some use different alloys to get special characteristics, but most just copy what Telma does.

I know of one manufacturer that offers anti magnetic SS frames for use around truck and bus electronics.

But, as I said in an earler post, they are the big kid on the block, and keep the competition small and struggling.

I hear that their parent company wants to unload them real bad. But I haven't independently verified it.

rmw