Faking a proportional valve

[Tinine]

Post #1

Anyone ever PWMd the coils of a bang-bang valve? I tried it recently so that I could reduce machine setup-time by eliminating the manual flow controls.

At no point have I suggested this as a substitute for a high-performance, zero-overlap servo-valve under PID control.


This is simply flow regulation. Many applications utilize a comparatively expensive Proportional valve to vary the speed of a hydraulic cylinder after some form of trip-point. The trip-point could be time, a limit-switch signal, analog/digital feedback, etc.


In my particular case, I would, ideally prefer to upgrade all actuators to full servo control but the client can't justify the extra-cost.
Because manual adjustments on an otherwise fully-automatic machine drives me nuts, as part of my control-system retrofit, I'm gifting a few enhancements to the client.


Instead of simply driving an SSR output fully on or off, I will use a simple H-Bridge driver and vary the coil-current via PWM.
This simply results in variable spool displacement. One can demonstrate this by manually pushing the spool using something like a small hex key/wrench.
Manual flow controls to be removed or left fully open.



Providing programmable parameters on the HMI is a trivial task.




Craig

 

[Tinine]

"Pipe Benders" get pretty sophisticated and expensive






Craig

 

[Tom Jenkins]

You could look at a simple motorized actuator on a ball valve instead of a spool. You monitor the flow and incrementally hit open or close on the motor operator until the flow rate is OK. I've used this system on butterfly valves to modulate airflow. There are two tricks - one is a slow operator so the movement increments are small. The other is to allow a time delay between movements to allow the flow to stabilize.

A refinement is to have movements closed and open of slightly different durations. That minimizes hunting.

 

[Tinine]

You could look at a simple motorized actuator on a ball valve instead of a spool. You monitor the flow and incrementally hit open or close on the motor operator until the flow rate is OK. I've used this system on butterfly valves to modulate airflow. There are two tricks - one is a slow operator so the movement increments are small. The other is to allow a time delay between movements to allow the flow to stabilize.

A refinement is to have movements closed and open of slightly different durations. That minimizes hunting.

I have been thinking about a rotary valve coupled to a small servo-motor....just for giggles. Admittedly I haven't spent much time looking but such a valve (good for 140bar) must already exist(?).


Craig

 

[Peter Nachtwey]

"Pipe Benders" get pretty sophisticated and expensive
Craig

I agree with the comments from adaptivephenix. I would add it is moving too much mass too far to simply bend a pipe.

 

[PreLC]

In any serious tool-based production center(Stamping, molding, Die Casting(Ignore melting cost to keep comparison standardized), bending etc, the cost is as follows:

1)Machine Capex,
2)Tool capex,
3)Tool maintenance,
4)Valve maintenance, and loss of OEE due to valve troubleshooting(Valves don't have a health check usually, differential pressure measurement is tough with multi port multi fluid multi valve systems),
5)Energy cost for hydraulics.

If you tell me that I can spend 5-10% of the Machine Capex on encoded valves to get the right type of flow control, I wouldn't be the one penny pinching. If someone suggests that any non-traditional valves are being used in the machine to save the machine cost, I would get them in touch with Parker/Festo/Vickers and ask them to get consensus, then sell it to me as a customer. I want my maintenance team to deal with #4 as little as possible.

That's the thing, if these systems worked effectively, someone would be selling them.
Maybe Tinine it's time to hang the ol' controls hat and put on the shiny leather salesmen boots and get rich off this idea.


It's always cool to accomplish things on your own machines, and I support it 100%. But like all engineering questions, "Should I use PWM to influence the spool to open the port's orifice X% to accomplish flow control" can only be answered by saying "It depends".

 

[Tinine]

I agree with the comments from adaptivephenix.

adaptivephenix sounds well informed and knowledgeable...oh wait, adaptivephenix is me!


This is precisely what I do....take machines like this that die due to lack of support, throw away the 840D (quite common) and put my own system on there that has a robot-like teach-mode.


The original manufacturers propose that you just spend another $500K on a new machine. The unique thing about this industry is that very few are able to resurrect these things.


Regular CNC machining centers can be and are retrofitted by everybody and his brother.


Craig

 

[Tinine]

Maybe Tinine it's time to hang the ol' controls hat and put on the shiny leather salesmen boots and get rich off this idea.

I sell a rugged Android tablet + microcontroller for $20K
The knock-on from that is that I become the go-to for every other issue that they have ($1K/day + expenses) because everything must be a "software problem", right?
"Hey we have this hydraulic leak...it wasn't there before we changed "the software"

It's always cool to accomplish things on your own machines, and I support it 100%. But like all engineering questions, "Should I use PWM to influence the spool to open the port's orifice X% to accomplish flow control" can only be answered by saying "It depends".

It depends whether or not you like wrapping your valve with 20 X $100 bills, I guess.


Totally off topic but this whole industry reminds me of what happens in the Guitar world (my hobby)


Grab a couple of TL072 op-amps, a few resistors, capacitors, an aluminum box and with the original, published, open-source schematic, you have a "Klon Centaur". Twenty five bucks, tops.



But because of a few famous, influential guitarists swear that the original was "special" and that it sounds "buttery, creamy, smooth", etc.,


You still get this:

 

[Tinine]

All that garbage behind the dry-wall, held together with wire-nuts? Yeah well ignore that because it's out-of-sight. The "audiophile" will swear blind that his $2k+ 3ft power cord from the wall to his amplifier results in "greater spatial definition"




The BS is everywhere and folks just lap it up




Craig

 

[Peter Nachtwey]

The BS is everywhere and folks just lap it up
Craig

The problem is too many do. So many aren't aware. So many are told the same lies so many times they think it is the truth.

 

[Peter Nachtwey]

adaptivephenix sounds well informed and knowledgeable...oh wait, adaptivephenix is me!

I thought it might be because of the claim to have patent. I don't know what you have a patent for or whether has run out.

 

[Tinine]

I thought it might be because of the claim to have patent. I don't know what you have a patent for or whether has run out.

In the early 1990's, Tier-1 exhaust pipes became "emissions control"
They switched from cold-rolled steel to 409 stainless, sensors dotted all over the place and deformation of the bends
was no-longer acceptable (turbulence).

What was formerly a sewer-system for the ICE was now a critical part of tuning.

CNC tube benders were traditionally only 3-axis. Two servo motors for tube positioning and a servo-hydraulic (MOOG valve)
bending axis. The problem with the hydraulic axis was the transmission which involved chains, yokes, idler-rollers, etc. and with the increased demands of 409 stainless, these already high-maintenance components became a nightmare.

I had the idea of using a servo motor thru a helical bevel gearbox but at the time, I couldn't find a big enough servo.
Wandering round a trade-show, I came across Unico and they had this huge, crude-looking AC induction motor (like a typical
pump-motor) sitting on a stand but with a BEI encoder sticking out the back...what the hey?
Then I realised that it was actually servoing. This was the time that I first became aware of AC vector control (FOC).

Talked to Reliance who came up with a low inertia motor that was nominally 10HP but with the addition of a blower, I could drive it with a 40HP Unico Vector Drive.

Built a machine with a new bending head, incorporating a 55:1 Nord helical-bevel gearbox and the first test was heavy-wall,
3" diameter 409 stainless. Holy moly 180 degrees/sec and the darned thing didn't break a sweat.

Big auto were ecstatic and orders poured-in. We went further and closed the loops on pretty-much anything that moved which
facilitated fully automatic setup and the elimination of all screw/pressure adjustments.
We had the 100% electric version and the hybrid (some closed-loop hydraulics).

Filed for and received the patent.

The problem with 100% electric was not the huge cost difference but the fact that I would refuse to guarantee "more accurate parts".
Our servo hydraulics were just as repeatable (Sony Digiruler feedback, 0.01mm res) and cylinders can take punishment way
better than any gear transmission, roller-screw, etc.

The patent was almost immediately infringed and in my naivety, I expected some official to pop-up and shoot these guys down
but nope. All the patent does is reserve the right to sue....and it's pay-as-you-go.

Ended-up putting all serious competitors out of business anyway.


Craig