Saw Dyno

[wcorey]

Ok, I'm at least home from work and can focus a little more clearly on this.

As to bar mounts and pitches, what I've got covered is pretty much anything that will bolt up to
do25 (stihl), doo9 (lg husky) or ko95 (sm husky) mounts in 3/8, 3/8lp or .325.
A lot of latitude here since I'm blocking off the oiler holes and using m'cycle chain lube to keep the mess to a minimum.
So there are a lot of saws that will mount up directly with a few various stud spacers and without the need to line up lube ports.

I have two 35 pin rims, one in .325 and one that will handle both regular 3/8 and also 3/8lp. They slide on/off the driveshaft and are retained with an 'E' clip.
All three drive chains interchange on two bars, one a ko95 and the other a d009 milled out to d025 slot width.
Each bar has two sets of mounting holes depending on which rim is being used, both rims are the same pin count so the 3/8 is obviously larger dia.
Three screws on the bar, possible chain swap and the clip is basically all it takes to change from one setup to another.

Since the rims are so large, the chain wouldn't clear the clutch cover on many saws unless the saw was mounted 2 or more feet back from the driveshaft.
To deal with this I added an idler/tensioner wheel that is sprung vertically and adjusts horizontally on a t slot.
Most saws can be installed on the bar regardless of where the saws chain tensioner post is adjusted to and the idler wheel is then positioned to take up the slack.

 

[paragonbuilder]

I'm glad you started the thread, there are a lot of knowledgeable guys on here that can add to this build.

im interested to see a few built up saws go on this dyno, and would like to eventually send one of mine as well

We started this before OPE. You can imagine how that would have gone on AS...
Good group of guys here though that can add insight with everyone getting pissy about opinions. I'm glad it's out here now for the added knowledge base. And so others can build there own if they want. That is what this is all about.

 

[Woodpecker]

Be right back, need to find an electrical engineer to explain this to me...

No kidding! I'd like to think I'm smarter than a bag of hammers, but I sure don't feel like it after Reading the explanation of how it works. Cool as heck though!

 

[Keith Gandy]

Cost to get a saw on dyno?

 

[paragonbuilder]

Cost to get a saw on dyno?

Let us get done and confirmed accurate before any of that stuff. Our goal is to get others to build them too around the country. Shipping would probably be the biggest expense. Shame to ship it 2 ways for a minute on the Dyno...

 

[paragonbuilder]

And maybe before others build them we could ship one around to gtg's. Right from the beginning this was about doing this for the community to help us grow. Personally I want one in my shop so I can learn as I do mods, what works and what doesn't on different saws. I don't care if I make the strongest saw, but I want to know what I've accomplished. And seat of the pants is very subjective.


Sent from my iPhone using Tapatalk

 

[jmssaws]

And maybe before others build them we could ship one around to gtg's. Right from the beginning this was about doing this for the community to help us grow. Personally I want one in my shop so I can learn as I do mods, what works and what doesn't on different saws. I don't care if I make the strongest saw, but I want to know what I've accomplished. And seat of the pants is very subjective.


Sent from my iPhone using Tapatalk

I'm building a new shop this year and already have a spot planned for one.

Unless something is drastically changed they all look the same to me,one of them will help that.

 

[paragonbuilder]

I'm building a new shop this year and already have a spot planned for one.

Unless something is drastically changed they all look the same to me,one of them will help that.

Seriously you don't need one!

I think you've got them figured out!
Lol!
And I need time to catch you...

 

[malk315]

So we did some more testing today to see if some of the damping ideas would help with the bounciness in the torque readings from the load cell.

The videos below show you can actually read the torque (Yellow LEDs) and horsepower (Green LEDs) live while running the dyno, something that was impossible before due to bounciness (noise) in the load cell reading. Bill was trying to work the brake to hold it at a peak of HP in the second video. First video you can actually see it hit north of 6HP a brief period -- around 9K RPM is when this saw does this (at about 17 seconds in you can see 8750 RPMs with 6.46 on the green display HP and 3.87 ft-lbs on the yellow display (torque).

Here is how we got there:

First up was 5 lbs of weight on the torque arm. One possible reason this didn't help much is due to the current load cell attachment not being rigid -- there is a U shaped "loop" (looks like a stirrup) that goes under the load cell beam and it can hang below it and put no pressure on it and when the brake pulls up it puts pressure on it only when pulling up. This had been done to allow different materials to be inserted to help w/ isolating vibration etc.

Here is a 30 second graph (time on the X axis w/ RPM, Torque, HP plotted) and you can see the bounciness in the torque reading big time still (not useful really):



Late last week I modified the software to switch from sampling the load cell every 100ms on the same period the RPM is sampled to using it's own thread sampling every 5 ms or 200 times per second. I then added decimation filter support to remove the noise. The above data was taken without the decimation filter active to see if the weight being there did anything. The below graph is using a filter width of 20 samples and it really cleaned things up a lot (not smooth, but way better):



The above has 3 times where Bill brought the saw with the brake from about 12500 RPM on down. The torque graphs are way better and you can start to see curves in there (it will be easier once RPMs are on the horizontal instead of time) I converted some to RPM + torque and HP curves I'll share below.

Here is another graph where the filter is set to a width of 30. This helps quite a bit with the spikes and out of the testing we did seems to be the best setting for this. This graph below is the graph from the 2nd video above complete with the brake getting dumped at the end and you get almost 10 lbs torque for just moment at the tail end of the graph. (Yes this spike at the end belongs! Can you say piss rev and dump the clutch? It's pretty cool to follow along with the video pausing or frame advance and match the sound / sight of the LEDs with the graph below. Luckily the load cell didn't snap since it was taking essentially it's load limit during that spike of about 44 lbs (we have a 5:1 mechanical advantage from the 35 pin sprocket the saw drives). Yikes.



So now we're getting somewhere -- we can likely go back to a rigid mount for the load cell which will allow us easier zeroing out to account for the weight of the brake and take it's weight out of the torque reading. We had to do that by taking up the slack in the current "stirrup" setup -- otherwise the weight of the brake isn't accounted for.

So when you take the period from about 2 seconds in above to about 15 seconds in and plot those like a normal Torque/HP curve with RPM at the bottom you get something like this:



The trends in black in red above are much easier to visualize than previous attempts with all of the noise in the data. There is still some variance around the points at the same RPM but that is also way improved over trying to do this w/ the noisy data.

Here is another plot where I took the data from 22 seconds to 26 seconds from the same run above which includes the 6.4 HP peak at 25.3 seconds in:



I should mention too that we also tried 2 different brake rotors to see if there was any difference in vibration or grabbiness of the brake. There is still noticeable vibration -- Bill can fill in here better he was mentioning as RPMs came down there was some type of left / right oscillation going on.

Apologies I don't remember the exact saw model w/ it's mods. It is a Makita (Dolmar) with big bore setup that gets over 80CC I think but I don't remember! Bill can reply and mention the exact saw.

That's the latest -- hopefully I can have the auto data recording (instead of timed) working by the CT GTG along w/ instant graphs w/ RPM on the horizontal. That way you can just do a run pretty quickly with the brake (like 5 to 10 seconds) and look at it on the web interface, do it again etc. and then save off several times where you run it down with the brake.

Quick edit: I forgot to mention too that this shows another improvement in RPM resolution before we had to the nearest 600 RPMs and it's now to the nearest 250 RPMs. We went from 1 magnet to 6 magnets per revolution that the HES sensor picks up. Also the RPM is more instantaneous requiring less time to go by to get a reading by a factor of 2.5X.

We will keep at it.

 

[mdavlee]

Nice improvements. Looks much easier to read

 

[MustangMike]

Way over my head, but best of luck with it guys, looking forward to seeing it at GTG.

 

[Brush Ape]

Bill and Eric couple of nuts. You guys shouldn't have any trouble gettin' a piece of ass with that thing.

 

[malk315]

Bill and Eric couple of nuts. You guys shouldn't have any trouble gettin' a piece of ass with that thing.

LOL. What lady doesn't like a device that spins, vibrates, and may cause a little friction!

Sent from my XT1080 using Tapatalk

 

[MustangMike]

CAUTION .... DERAIL .... CAUTION .... DERAIL!!!

 

[malk315]

Going to use the dyno tomorrow late afternoon (Friday) the day before the CT GTG at Dan Paragon Builder's place.
I did a bunch more to the software to support automatic data collection based on a load going active and inactive along with easy click to look at graphs after that have RPM along the horizontal. The new parameters to support it along with a graph of what the auto collection would have taken had it been working this past Sunday for one of the passes Bill did with his Makita/Dolmar saw that looks really good:





Unlike Excel this graph doesn't care that the bottom axis isn't in order or if there are duplicate points.
Since this is a screenshot you can't see it, but when you hover the mouse over a point, it gives you the value and the RPM the value was taken at making it simple to know the exact numbers for any point w/o having to guess based on the Y axes.

On the live LED displays when an auto run starts capture it shows live on the Yellow display on the unused digits to the left the count of data points taken so at a glance you can tell while running the saw and hand brake each time a batch of data points is collected -- you can quickly do a few passes and then look with web interface how you did instantly.

Pretty happy with the above curve. Hopefully we'll see some decent ones tomorrow too.

 

[Tor R]

Your guys have done awesome work to put this together.
Looking forward to hear how it work out on the GTG

 

[Redfin]

Very nice work fellas!!!

 

[MustangMike]

Any reason the RPM range is lower than I would have expected?

 

[jmssaws]

I'm anxious to see a particular saw on it.
2 saws actually.

 

[paragonbuilder]

Any reason the RPM range is lower than I would have expected?

You mean peak rpm? Or where the power is?