I'm really hoping that Evil Aviator will see this topic, since it is he who originally got me thinking about this when he randomly mentioned it in a post.

Does anyone have experience with turbine tachometers? What is the advantage to knowing the turbine speed? I would assume that it would make it easy to detect a stall condition, or compressor surge, but what else?

I've searched around a little on Google, but have found very little information. Most of it was related to aircraft, and as you can expect, aircraft level instruments are very expensive. Is there anyone that produces an automotive turbine tach?

Thanks!

 

Not sure how applicable this is: but they make optical tachs for model aircraft that read quite high-rpm by counting the number of blades that pass between the sensor and a light source. The principal is pretty simple because they have been around for years. You could try something similar.
Grant

 

I have seen this gauge before on the internet....but I cant find it...

Sure is cool but you have to have the optical sensor in front of the spinning compressor blades...hope it is sturdy.

 

Evil I am cretain has all the actual info.. but weve been looking for a affordable method for the 20B's turbo for a very long time.
IIRC there is a tach available from Aircraft Spruce that was a pretty good fit.
Then again... it may have been via a high speed optical sensor used with a digital meter of some sort... ala multi meter.. but it HAS been a while. I am 99 percent positive he has the info in his volumes of collected stuff.

 

It's just an optical sensor that can handle very high frequencies.
I dunno if it's actually a "turbine" tach, but the ones I've seen are mounted on the compressor side.
It just triggers a pulse when one of the compressor fins go spinning by.
Divide by the number of total fins, and viola...your get "RPM" for the turbo...


-Ted

 

Making one is not the issue...I'm mainly just wondering if it is worth it. What can this gauge tell an operator about the state of the car besides what I have already mentioned?

As far as construction, any suitably fast phototransistor mounted into the compressor housing would do the job. Shine IR light onto the fins of the compressor, but paint the tip of one blade black. Suitably fast op-amps have been available for years. The output of that simple circuit would be a squarewave showing one pulse per rotation of the compressor. could then be divided down for feed to a standard AutoMeter tach, sent to a micro for further processing, etc....

 

Yes, that is true.
Finding an optical sensor and emitter is not that big of a deal.
Designing a simple circuit for a counter should be a piece of cake for an EE freshman.

Someone showed me that eBay link to the "box" that was supposedly like $2,000 brand new!
Yes, it is overpriced for what it does.

What it does is allow you to check the turbo speed relative to power.
Obviously, if you're making more power with less turbo speed, you're tuning is better.
You can also check to see if your turbo speed matches the compressor maps for any given turbo / compressor.


-Ted

 

Yep one of these devices could have come in handy with Red-Rx7's 20b t72. It would have shown that his turbo was quicky running out of it's efficiancy range due to excessive compressor shaft speed.

 

http://www.monarchinstrument.com/Pocket%20Laser.htm

 

Digging the idea, but whats going to happen to your optical sensors and your electrical bits when the turbo gets that lovely red glow after a hard run?

 

The optical sensor is pointing at the compressor wheel not the turbine wheel.

James

 

$2000...wow. I did a quick parts lookup yesterday, and it looks like total cost for the circuit will be under $20. I guess it will be a small winter project.

That pocket laser tach is very cool, and not very expensive either. Might be an option.

 

From what i know F1 Teams of back in the day used Magnetic pickups with the magnet embedded in the Shaft. This is probably not feasable for most us though.

 

I'd recommend making sure you get military/automotive grade parts for the IR emitter and phototransistor. I bet even the intake side of the compressor gets pretty hot.

Another trick is to modulate your light source at a specfic frequency. This should boost your SNR. Might also want to use a visible or UV emitter. Seems like you have more heat generated radiaton to worry about than solar radiaton.

I did a visocsity measurement lab in college that had trouble with false triggers. This was later resolved by freqency modulating the light source, with the appropiate filter on the pickup side.

Just my two cents.

Sounds like a neat project.

 

Look at how the guys who make jet engines out of turbos do it. They tend to rig up some cheap, but effective solutions.

I believe one solution involved painting the compressor side nut black, and painting one side white. They then ran a fiber optic cable down into the housing, mounted it securely, and used it to catch the reflected pulses. Something like that.

I'd be interested to run the output of the turbine tach into a datalogger, so you could plot it against RPM & throttle opening & manifold pressure & such - get some good tuning data, I'd think. And, at the very least, another nifty gauge in the car that read really high numbers.

-=Russ=-

 

That sounds like a really elegant soluation!
IIRC real fiber is glass, so temps aren't an issue.

 

To certify a turbocharger in aircraft applications it must be tested throughout its flight envelope to ensure overspeed is prevented. Here's a method Garrett/Allied Signal uses for this application.

The turbo is disassembled, and a flat machined on the shaft. The rotating assembly must then be re-balanced. The center housing is drilled and tapped to suit an industrial VR sensor, such that the VR sensor is aligned radially with the centreline of the shaft, and axially with the newly machined flat on the shaft. Once assembled and adjusted properly, the VR sensor provides a pulse each revolution as the flat passes the sensor.

This method is robust, easy to adjust, and there is no risk of dropping bits into the rotating assembly. The other advantage is that the instrumentation stays in the center housing, allowing easy turbine/compressor housing swaps during the test phase. Applications I'm aware of were always oil-only center sections with plain bearings. Suitability of this method to use in water-cooled/ball bearing turbos would need to be reviewed.

As for 'why' you want to know rpm, it's a relatively easy parameter to measure that allow evaluation of the operating point on the compressor map. Measuring pressure ratio is also easy, but measuring mass/volume flow with the same accuracy as the other two is not nearly so simple. Besides, when testing, you can never have to much data!

 

That is precisely what I had in mind. Except that I was just planning on painting a small black mark on one of the blades of the compressor.

 

http://www.me-us.com/turbo/

http://www.acam.de/index.php?id=15&L=0

But the output is in volts...what do you do with it then? Datalog the volts then convert it with an excel spread sheet?

James

Edit - Thanks 1987rx7guy and his searching skills

 

Holset used these on some of there current production turbos for heavy duty applications. I cannot remember if their medium duty (cummins turbo diesel p/u's)have the sensors. The sensor plugs into the bearing housing and the shaft has a flat spot ground on it. you might be able to find a pn through Volvo or another one of Holsets customers.

Justin

 

strobotach