Saw Dyno

[malk315]

Hydraulic Dyno testing was yesterday. Obligatory pictures and video:

 

[RI Chevy]

Nice job guys. Keep up the great work. [emoji106]

 

[malk315]

Raw load cell data comparison -- From the motorcycle dyno we had this 3 seconds of data where the torque was ranging from 0 to 25 lbs -- a large range compared to raw data on the new one which is next:

Hydraulic dyno raw torque measurements 400 samples per second (like above graph was taken), but many more seconds here:



In the above it's only from 0 to 4 lbs the entire time. This has essentially the same reduction from saw to pump of 5:1 where saw rim was 7 pin to jack shaft 7 pin that had a roller chain sprocket with 11 pins out to a 54 pin on another shaft to drive the pump.

With just a 10 point decimation filter it already cleans up better than motorcycle setup. Motorcyle setup took 60 point filter to calm things down

 

[malk315]

The new dyno has the RPM sensing at 1:1 instead of on the 5:1 reduced shaft and Bill has 16 magnets on there for 16 pulses per revolution. The old one had RPM resolution to the nearest 250 RPMS (it would have been 50 RPMs if 1:1 instead of 5:1 with 6 magnets on old setup). The new one has better than nearest 40 RPM resolution a more than 5:1 improvement.

The new dyno pump creates a load you can't get rid of such that my stock 372XP X-torq saw maxes out at 11,000 RPMs or so running the pump with little to no flow restriction by the valve. It translates to a constant torque going in, but since the valve isn't applying a decent load, you get this flatness of the graph for a while at 10000 - 11000 RPM depend on how much you loosen the valve to start your run and then the extreme right-hand side of the graph is where it gets more interesting. I will need to adjust the software to be more picky about where it graphs to avoid the mostly boring stuff up to the extreme right of the graphs below.

We were trying different filters here -- 10 point through 60 point in increments of 10. 30 point seemed to be plenty to get useful data.

I think the key here is the 372 is showing a rather consistent 4.0 HP or so at 9K RPM or so. Much more consistent than the motorcycle one IMHO.
In the graphs below when you look at peak area it's always 4 HP at around 9K.

I know that's a bit low for a stock 372, I can't explain why it's low we may have something wrong or haven't factored in something. We were fighting daylight to get the runs taken. What I'm happy about is the consistency -- all w/in 0.1 to 0.2 HP of one another instead of much larger changes.

Wondering if the pump is possibly cavitating or if there is some kind of blow by going on.
Here are the graphs given the older start/stop strategy which isn't the best for the new setup.

I'm going to include RPM on bottom (which screws the scales on either side for torque and HP) and then the ones with the RPM as an axis and time on the bottom which forces torque and HP to use the same scale and then you can see them cross at 5252 RPM like they should:



Fine resolution on the RPM blue line below. Black torque and Green HP cross at 5252 RPM at the extreme right of the graph which with hydraulic dyno extrreme right of the graph is the most interesting part.






We get the peak HP at 4.0 HP or so again here. Across several runs they are all w/in 0.1 HP of 4.0 on this setup.

 

[malk315]

A couple more graph screenshots I took -- this run #488 we were using smaller filter on the torque you can see it... But if you imagine a trend line it's still 4.0 HP at around 9K RPM... very consistent!



Below looking at the RPM plotted the beginning has a couple of throttle blips that we caught and then it goes up to just about 11000 and the torque and HP lines go up on an even slope as the valve is cranked in and 9000 on down is the most interesting part on extreme right of the graph -- you have the usual torque and HP curves crossing at 5252 RPM. I will take the raw data from these graphs an chop out the left hand data so the extreme right widens out more and re-plot them the way a car dyno would make it look. Will also try binning the torque points to nearest 250 and 500 RPM bins like the guys do on manual dyno measurements and I'll bet those "cleaned up" graphs will look really consistent and much like we expect for torque/ HP curves with RPMs increasing instead of decreasing along the bottom. Once I figure out what will look best I will make the software work that way too.

 

[Lee H]

You and Bill have done a great job on this project. I see a lot of hr's
spent trying to get it right. If and when you two get it all sorted out
i would love to be able to put one of my big saws on it for chits.


Lee

 

[malk315]

You and Bill have done a great job on this project. I see a lot of hr's
spent trying to get it right. If and when you two get it all sorted out
i would love to be able to put one of my big saws on it for chits.


Lee

You bet Lee! I think we can get there -- Bill may want to try a different valve setup to try and give us a bit more control. I will change the software to bin the torque / HP readings into RPM ranges and average them so we get an automated graph that looks like what Chad and Dieselshawn get on their dynos by hand i.e. every 500 or 250 RPMs we can apply points taken into those bins and it won't matter how long we "dwell" with the valve at various RPMs. They should be all relatively close together and when averaged will get rid of any remaining graph ugliness hopefully.

More to do... wish I could take a month off from work to hack around on this thing even more.

Will shoot to have something usable by Dan @paragonbuilder spring GTG.

I think I've been too distracted by the Tennessee GTG thread -- man that thing is like a realtime election chat feed... pages and pages of posts.
Gonna mark my Calendar for 2017 Tennessee GTG. Doc Al @drf256 talking about renting a bus if I remember right?

 

[RI Chevy]

Lol. I hear you Eric. I unsubbed. Lol

 

[malk315]

By the way I'm laughing watching my video at the part:

Me: "Get a little footage of it"
Wife: "Honey?"
Me: "Before we start"
Wife: "Honey!" (ignoring didn't work)
Me: "Hold on I'm filmin somthin"
Wife "Can I have the extension cord"
Me: "Yeah I can't give it to ya til were done".

Wife wanted the huge extension cord that powers the dyno electronics for the outdoor Christmas lights my wife and kids were setting up... oh priorities.

 

[RI Chevy]

Lol

 

[94BULLITT]

Very cool! I like how compact it is.

 

[Mastermind]

You guys are good.

 

[Lee H]

You bet Lee! I think we can get there -- Bill may want to try a different valve setup to try and give us a bit more control. I will change the software to bin the torque / HP readings into RPM ranges and average them so we get an automated graph that looks like what Chad and Dieselshawn get on their dynos by hand i.e. every 500 or 250 RPMs we can apply points taken into those bins and it won't matter how long we "dwell" with the valve at various RPMs. They should be all relatively close together and when averaged will get rid of any remaining graph ugliness hopefully.

More to do... wish I could take a month off from work to hack around on this thing even more.

Will shoot to have something usable by Dan @paragonbuilder spring GTG.

I think I've been too distracted by the Tennessee GTG thread -- man that thing is like a realtime election chat feed... pages and pages of posts.
Gonna mark my Calendar for 2017 Tennessee GTG. Doc Al @drf256 talking about renting a bus if I remember right?

Can you speak english,
Feel like i'm on the phone with some tech line.

 

[srcarr52]

You're going to want a bigger reservoir for dwell time of the hydro fluid. The throttle valve causes a lot of cavitation of the fluid and you'll need the dwell time for the bubbles to come out. The reservoir on my hydro kart dyno is 15 gallon.

 

[wcorey]

You're going to want a bigger reservoir for dwell time of the hydro fluid. The throttle valve causes a lot of cavitation of the fluid and you'll need the dwell time for the bubbles to come out. The reservoir on my hydro kart dyno is 15 gallon.

I know what you're saying but I figured if an automotive power steering circuit can get away with it...
Not that there aren't differences, this isn't going through a steering box/rack but the valve here plays more or less the same role.
Do you think cavitation bubbles could be causing some of the oscillation in the load cell output?

 

[malk315]

You're going to want a bigger reservoir for dwell time of the hydro fluid. The throttle valve causes a lot of cavitation of the fluid and you'll need the dwell time for the bubbles to come out. The reservoir on my hydro kart dyno is 15 gallon.

Hey Shaun -- thanks for this. Bill if we can add a big reservoir we can check 400 sample per second load cell data to look for improvements there.

Shaun can you take a look at the raw load cell data from above for improved filtering? Or should we wait til we try a reservoir and see if it smooths out on the raw readings even more. I'm liking that the range of values is much less on the hydraulic setup than the brake dyno setup. 0-4 versus 0-25 swath is a big difference. If you want to look at the .csv file let me know and I'll send it along.

 

[srcarr52]

Hey Shaun -- thanks for this. Bill if we can add a big reservoir we can check 400 sample per second load cell data to look for improvements there.

Shaun can you take a look at the raw load cell data from above for improved filtering? Or should we wait til we try a reservoir and see if it smooths out on the raw readings even more. I'm liking that the range of values is much less on the hydraulic setup than the brake dyno setup. 0-4 versus 0-25 swath is a big difference. If you want to look at the .csv file let me know and I'll send it along.

The hydro setup is a huge improvement already. Frequency filtering instead of decimation may now work better.

I also think increasing your line size and throttle valve will help, especially with the max RPM issue. On the dyno I have it's 3/4" line to the throttle and 1" return. No need for a cooler when you have a large enough reservoir.

 

[Red97]

I think a beer keg resivor would be pretty sweet. Attached to a cart, along with the readout box. sturdy compact unit.

Thank you guys for sharing all of this.

 

[malk315]

The hydro setup is a huge improvement already. Frequency filtering instead of decimation may now work better.

I also think increasing your line size and throttle valve will help, especially with the max RPM issue. On the dyno I have it's 3/4" line to the throttle and 1" return. No need for a cooler when you have a large enough reservoir.

Frequency filtering is not my area of expertise. Any chance you could help me out there?

I think though we can wait and see if mechanical improvements help out first -- would really cool if we didn't need to filter and just bin the datapoints collected into various RPM ranges and plot from low to high w/ points that fit in each range averaged. A plot of resulting points may end up looking really nice. I can make the software do it automatically (like folks have done manually by reading a scale at various RPMS dialed in with the value) and make the range of each bin adjustable -- to try 50, 100, 200, 250 rpm ranges. I've got plenty of resolution on both RPM and load cell to do this. I'm going to also make the RPM display interval independent of RPM collection interval -- the display updating every 10th second is good but I can collect RPMs more often then that now that I have more pulses per rev and no reduction ratio robbing of resolution.

Also Shaun I'm very happy that the 16 pulses per rev on the hardware encoder interface on the beaglebone computer keeps up no problem w/ shaft spinning at saw RPM rate. My RC filter for noise is good through ~15000 hz frequency so no problem there.

 

[malk315]

I think a beer keg resivor would be pretty sweet. Attached to a cart, along with the readout box. sturdy compact unit.

Thank you guys for sharing all of this.

Yup -- bring your keg of beer and collection of saws to dyno. When keg is empty we'll fill w/ fluid and dyno all of the saws as a "thank you" for the beer