Saw Dyno

[czar800]

Y'all realize that when we get hard data, and realize that this is all pointless, Randy and the others that rely on porting for their bread and butter, are going to be out of a job...

He has been warning us

 

[czar800]

Very cool stuff!! Put a tip can next to it at the gtg!

 

[srcarr52]

Yes -- that's what we're seeing. Next run I'm going to try sampling the load cell 10x more often and using 1/10th of each sample (decimation filter) to get it more under control. Bill already did a mechanical change that appears to have improved from a 2 to 4 ft-lb swath to a 1 ft-lb swath with the latest run.

If we do change to some other method for the power to push on a load cell that is smoother, the electronics can work the same -- so that's good for what I'm working on.

My day job involves linear synchronous motors and of course I've shared with some of the fellow nerds here the dyno project.
We were looking at eddy current brakes -- those would be ideal since there is no contact, should be perfect smooth power transition.
Some that go on exercise bikes are cheap -- one we looked at has a sort of cable to put the brake in a position that makes the eddy currents happen and create the braking force... with no friction contact.

Always open for ideas -- that's how we've gotten this far! Thanks.

If I would have known what you all were up to I would have told you to not go down the friction brake road. For a cheap build without a lot of electronics a hydro pump and throttle valve work pretty well. Eddy current needs some extra electrical work to be able to control the load. I have heard of small eddy current dynos built with an alternator.

A decimation filter like that, (under sampling and averaging), can only get you so far. A true low pass filter is much more mathematically complicated to do to a digital signal but you can always rig up an electrical low pass to the signal. It's just an opp app, couple of resistors and a cap. I would think a low pass of 5Hz or so should get you where you need.

Something else you can do to reduce noise is to add mass to the torque arm. It's a dirty trick that is often used in the NVH world but it's really effective.

 

[malk315]

If I would have known what you all were up to I would have told you to not go down the friction brake road. For a cheap build without a lot of electronics a hydro pump and throttle valve work pretty well. Eddy current needs some extra electrical work to be able to control the load. I have heard of small eddy current dynos built with an alternator.

A decimation filter like that, (under sampling and averaging), can only get you so far. A true low pass filter is much more mathematically complicated to do to a digital signal but you can always rig up an electrical low pass to the signal. It's just an opp app, couple of resistors and a cap. I would think a low pass of 5Hz or so should get you where you need.

Something else you can do to reduce noise is to add mass to the torque arm. It's a dirty trick that is often used in the NVH world but it's really effective.

Ahh -- I see. Low pass filter on the electrical part of the load cell. Great idea. I've got squat for room on the breadboard but I'll see if I can squeeze something in.

Adding mass to the torque arm should be easy -- but to cancel out gravity we'd have to rotate the whole thing sideways.
I can still zero the load cell to get rid of the weight when it's upright... but we'd be limited as to how much mass we could add... what's interesting is the torque arm pull upwards against gravity, so we could weigh it down and then pull back against it...

Thanks a bunch for the ideas.

 

[paragonbuilder]

If I would have known what you all were up to I would have told you to not go down the friction brake road. For a cheap build without a lot of electronics a hydro pump and throttle valve work pretty well. Eddy current needs some extra electrical work to be able to control the load. I have heard of small eddy current dynos built with an alternator.

A decimation filter like that, (under sampling and averaging), can only get you so far. A true low pass filter is much more mathematically complicated to do to a digital signal but you can always rig up an electrical low pass to the signal. It's just an opp app, couple of resistors and a cap. I would think a low pass of 5Hz or so should get you where you need.

Something else you can do to reduce noise is to add mass to the torque arm. It's a dirty trick that is often used in the NVH world but it's really effective.

We went back and forth about it. I think on a manual setup where you are reading a scale like Chad had it would work ok. I could be wrong.
Adding mass sounds like a great idea!

 

[paragonbuilder]

Ahh -- I see. Low pass filter on the electrical part of the load cell. Great idea. I've got squat for room on the breadboard but I'll see if I can squeeze something in.

Adding mass to the torque arm should be easy -- but to cancel out gravity we'd have to rotate the whole thing sideways.
I can still zero the load cell to get rid of the weight when it's upright... but we'd be limited as to how much mass we could add... what's interesting is the torque arm pull upwards against gravity, so we could weigh it down and then pull back against it...

Thanks a bunch for the ideas.

That makes sense eric. If the load cell was sideways would be better, or else it creates more parasitic loss. Hmm. Thinking...

 

[srcarr52]

We went back and forth about it. I think on a manual setup where you are reading a scale like Chad had it would work ok. I could be wrong.
Adding mass sounds like a great idea!

I have a dyno of similar design to Chads but it's for kart engines. Manual reading sucks and it requires you to hold stead state long enough to read the gauges.

Ahh -- I see. Low pass filter on the electrical part of the load cell. Great idea. I've got squat for room on the breadboard but I'll see if I can squeeze something in.

Adding mass to the torque arm should be easy -- but to cancel out gravity we'd have to rotate the whole thing sideways.
I can still zero the load cell to get rid of the weight when it's upright... but we'd be limited as to how much mass we could add... what's interesting is the torque arm pull upwards against gravity, so we could weigh it down and then pull back against it...

Thanks a bunch for the ideas.

All you need to do is zero the load cell with the mass on the torque arm, the rate of change of the force is still going to be the same. Yes, you would be limited by the max load of your load cell, good thing gravity and the saw act in opposite directions in your design.

 

[malk315]

I have a dyno of similar design to Chads but it's for kart engines. Manual reading sucks and it requires you to hold stead state long enough to read the gauges.



All you need to do is zero the load cell with the mass on the torque arm, the rate of change of the force is still going to be the same. Yes, you would be limited by the max load of your load cell, good thing gravity and the saw act in opposite directions in your design.

Yup -- we're thinking alike.
Bill when you read this perhaps you could see about a way to add some weight to the torque arm. Might have to go back to the more rigid connection so it just damps things out. We can zero it when idle and see how weight does.

If I have the dyno at my work we can put a scope on the load cell and see what it's doing and come up with a good filter there too. I was just chatting with one our guys here who knows the electrical stuff better than I do. I'm a software guy by trade but like to muck with the hardware too.

 

[Deets066]

Yes -- that's what we're seeing. Next run I'm going to try sampling the load cell 10x more often and using 1/10th of each sample (decimation filter) to get it more under control. Bill already did a mechanical change that appears to have improved from a 2 to 4 ft-lb swath to a 1 ft-lb swath with the latest run.

If we do change to some other method for the power to push on a load cell that is smoother, the electronics can work the same -- so that's good for what I'm working on.

My day job involves linear synchronous motors and of course I've shared with some of the fellow nerds here the dyno project.
We were looking at eddy current brakes -- those would be ideal since there is no contact, should be perfect smooth power transition.
Some that go on exercise bikes are cheap -- one we looked at has a sort of cable to put the brake in a position that makes the eddy currents happen and create the braking force... with no friction contact.

Always open for ideas -- that's how we've gotten this far! Thanks.

They use eddy brakes on older overhead cranes before they had inverter drives.

Any way of fitting an inverter drive and controlling with the computer? You cand change ramp up and down times any way you wanted

 

[Red97]

Sure do appreciate you fellas sharing.

Who made the sprocket? I'm gonna be needing one in the 84 tooth range..

 

[paragonbuilder]

Sure do appreciate you fellas sharing.

Who made the sprocket? I'm gonna be needing one in the 84 tooth range..

@wcorey did all the fab work. He is a talented man!

 

[Deets066]

They use eddy brakes on older overhead cranes before they had inverter drives.

Any way of fitting an inverter drive and controlling with the computer? You cand change ramp up and down times any way you wanted

Inverter drives usually control a motor, not sure how it could work with this setup, sure would be cool though

 

[idiotwithasaw]

That looks awesome. However one question I have that I haven't seen asked or answered. What about pitches? It appears that you are doing all initial testing on 3/8, will you be able to set it up to read .325 or even lo pro picco? It would be nice to see differences in torque between the different pitches and sprocket sizes.

 

[wcorey]

Hey all,
Following along with the various related threads today but still at work and has been too hectic for any long comentary.
Some good thoughts here, I like the idea of adding mass to the torque arm, very easy to do.
Still need to try a few other things first, we haven't had much time with the data collection combined with the mechanical hardware, really only minutes of actual run time.
Eric needs the dyno and I need the electronics but we only have one of each for now so it's limited what can be tested.

Eric was excited to go public with a thread and I don't mind he did but for me this thread is a bit premature. Was my fault anyway as I let the cat out o the bag with the hardware at the NY gtg.
I still have a lot of options left untried and so far very limited data to draw conclusions or opinions from.
Way too much of my input at this point would be just speculation, next week maybe not so much...

The construction of the dyno is purposely very modular, to allow for the many changes I expect to happen as we progress. I can and probably will remove the brake at some point and run some alternative loads.
The brake was first up simply because I already had most of the components on hand, though I ended up buying some cheap stuff so as to not 'waste' my good stuff. I don't think I have more than $30 into the whole thing, so far...
Since then I've found a large power steering pump that looks promising and a nice impeller that looks like a good base to construct a water pump from. Every design has it's tradeoffs though...

I'm looking forward to getting together with Eric this weekend for the next round and I'm sure we'll have a better idea of where we need to go from there.

 

[wcorey]

That looks awesome. However one question I have that I haven't seen asked or answered. What about pitches? It appears that you are doing all initial testing on 3/8, will you be able to set it up to read .325 or even lo pro picco? It would be nice to see differences in torque between the different pitches and sprocket sizes.

Not to worry, its all under control, there's an awful lot that has yet to be covered here.
Later...

 

[malk315]

Inverter drives usually control a motor, not sure how it could work with this setup, sure would be cool though

My day job is in the linear motor business and we use inverters to drive them. You can certainly drive them for braking force -- as you regen power you have to do something with it -- we usually dump it in adjacent motors with no magnet arrays over them. So a rotary motor could be driven to provide braking force -- servo drives for rotary motors we've used have a brake feature. I'm sure something slick could be created but could be expensive. Not my area of expertise as I work on next layer up control software.

 

[paragonbuilder]

That looks awesome. However one question I have that I haven't seen asked or answered. What about pitches? It appears that you are doing all initial testing on 3/8, will you be able to set it up to read .325 or even lo pro picco? It would be nice to see differences in torque between the different pitches and sprocket sizes.

If you are talking about trying different setups on the same saw, it shouldn't matter. We are measuring motor hp and torque. If you change the sprocket on the saw you need to change the settings in the software.
But eventually all sizes will be able to be used.

 

[malk315]

Guys if I don't answer questions right away don't think I'm dissing you I don't always stay caught up on the forums.
Most of the questions have been mechanical (sorry Bill!) but a couple have been asking detailed electronics stuff (and making great suggestions) so thanks! I'll do my best to answer in good time.
I like to tinker and learn things inside and out the best I can (probably how a lot of saw builders are and why we have such great ones on here) and I'm willing to keep trying things and moving forward til out of ideas... could be a long haul but that's OK.

Thanks again!

 

[Deets066]

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

 

[ABarrick]

Just when I'm starting to feel kinda smart, you post this and now I feel dumb again.

Really cool looking fab work. Wish I understood the electronics portion more but great job either way. Makes me want to drive to CT for a gtg now.