A plastic valve

Some considerable time ago, and quite by chance, I found myself designing and making a miniature five-ported hydraulic valve for use in medical research.

It was machined from solid Perspex (PMM) and measured about 26mm long by 22mm wide by 16mm deep. It was to be secured to a microscope stage operating at fairly high magnification, and hence had to be remotely controlled to avoid generating vibration. I chose to make the spool from Teflon (PTFE), which had to seal completely, two fluids being switched by the valve as the fluids passed through it. In other words, there must be no leakage.

The main bore of about 9mm diameter had to be machined and reamed with the minimal amount of frictional heat, so that final lapping (essential for accuracy and finish), didn’t cause crazing. Some liquids can be seen to induce stress-cracking from inbuilt stress, (a bit like happens when a high density polyethylene injection moulding comes into contact with detergent solutions). This crazing of the PMM was initially found to be a significant setback, until I developed a satisfactory low heat machining solution.

Once the word got out, I finished up making and selling several dozen of these individually manufactured valves to various research groups around the world.

My question is :-

To obtain a positive seal between the inside of the PMM and the PTFE spool, while bearing in mind that PTFE does creep, I made the centre section of the PTFE spools a small amount larger in diameter, ie. about 0.008mm (0.0003"). This was all I was prepared to trust from my lathe in a semi-production sense. The unfilled PTFE was ideal in terms of its low friction characteristics, especially that of `stick’ friction. Both materials (PTFE and PMM) had to be relatively inert to the fluids being processed.

In response to PekkaNF’s comments about machining PTFE, it's a material which is not easy to grip. Unlike POM (polyacetal, Delrin,) which is more rigid, and swarf normally comes away cleanly (unless static builds up and pulls the swarf onto the workpiece), unfilled PTFE is soft and slippery. Its creep resistance is very low too, so it becomes necessary to surround the workpiece with metal as a means of reducing movement (creep). At the time, I had quick release collets, so gripping the PTFE bar-stock was not an issue.

Another aspect which came to my notice in terms of the performance of the valve, was that at 19 degrees C, PTFE goes through a phase change and sharply increases in volume. This had to be taken into consideration for valves specified for use in applications where the operating temperatures were above 0 degs C, but below 19 degs C.

For these special lower temperature versions, I made the diameter of the PTFE spools a small amount larger to accommodate this drop in volume. I also had to supply special instructions for when the valve was not in use. Researchers could either keep the valves in the fridge or strip them down to remove the spool. Storing the separate parts at room temperature was not a problem, although an important part of preparation for use was to reduce their temperature to below 19 degrees, and to keep them below 19C while in use.

Returning to de-burring, I think Weldsol’s comments about using a sacrificial bar has merit, although I needed to make quite a few of these valves, so it would be necessary to jig the bar to accurately locate the hole centres and avoid significant (bar) wastage. This wouldn’t have been too much of an issue since I was already drilling the holes with a drilling jig. However, I do wonder about what would happen if the drill wandered slightly.

It doesn't want to let go! 

Regards to all,

Gentlemen,

Thank you for your compliments, I’m only pleased to have somewhere to divulge some of my `tricks’.

Before I disclose my method of de-burring the 1.0mm holes into the body of the PMM (Perspex) valves, I need to make a couple of comments about the more recent postings.

I fully agree with Pekka regarding the machining and use of unfilled PTFE (Teflon). However, once its basic properties like creep, coefficient of friction, and inertness are understood, these can be dealt with as required.

Pekka, your comments have aroused my curiosity. I suspect that you have worked in a research environment where these valves (or similar) were in use. I designed them specifically for one task in a biological research laboratory. From your other comments, I have to say that I also developed and made several other pieces of laboratory equipment, none the least being an infusion pump which would physically accept standard syringes from 5ml to 50 ml.

Stepper motors were indeed used as the motive force, while the engagement and re-setting of the pump and syringe was achieved using a gearing-down lead-screw system not unlike a lathe split-nut arrangement. As I recall, with a wide range of speeds and syringes, discharge amounts could range from a millilitre or two over several hours, through to a relatively quick discharge of 50ml in just a few minutes.

Another very serious semi-production task I developed, also involved machining elastomers. This was where the kitchen freezer became an import part of the equipment. It was of course, necessary to push the frozen food to one side, to make room for the items to be `frozen’.

Coping with the high `discharge’ rate of the polyurethane swarf was perhaps the hardest part of this production line, until my son and I decided to let the swarf wrap itself around the wheels and remove it later. A sharp Stanley knife worked wonders.

As for the PTFE spools, and to pick up on the comments from Pat, Hansrudolf, and Andrew. Using the 1ml syringes as the `power’ source as mentioned before, the spool was caused to slide a mere 2mm, so that either one of two fluids could be switched rapidly as they were directed into the specimen bath. A material combination with minimal stick-friction was required. The best materials I had access to, and which were suitable for my modest home workshop were polymethyl methacrylate (PMM - better known as Perspex), and polytetrafluoroethylene (PTFE - better known as Teflon). Together, and with the degree of fit I had determined as suitable, these two materials produced virtually no stick friction, and provided a positive seal between the adjacent fluid channels.

Since you asked Hansrudolf, all of my work in making these valves was carried out on my Myford ML7, and a nerve-wracking Chinese mill/drill. It wasn’t (and dare I suggest still isn’t), possible to use a micrometer to measure the central diameter of the spools to the level of accuracy I found necessary, especially since this middle section (web) of the spool was only 1.5mm wide.

At this scale, PTFE is too soft and slippery to get any kind of sensible reading. What I did was to use Go/NoGo gauging and selective assembly. For repeatability, the Myford was good enough for me to leave the cross-slide locked, and make final surfacing cuts without disturbing anything but the saddle.

Getting back to the original Thread, and why de-burring was so important. The minimal interference fit of the central section of the spool and the valve body, was still such as to cause the PTFE to (dare I say) exude into the 1mm port holes. Therefore, any kind of sharp edge was likely to interfere with, and even shear off, a tiny sliver of PTFE from the spool. Since some of the research involved measured millivolts, any liquid bridging the central spool web was likely to interfere with these low voltage measurements.

This posting has growed like Topsy, so I’ll add another posting to describe how I de-burred the holes.

Regards to all,

Sam

This posting could be titled - How I de-burred the tiny holes in the PMM valve bodies.

Just to back-track a moment, the inside bore of the PMM valves had a honed and highly polished finish. That was eventually my minimum standard. Where the five 1.0mm holes broke through however, the best result was a sharp corner, while the worst was a burr. Clearly, the PTFE spool would be damaged if this condition prevailed.

The solution was quite a simple idea really, and one which Pekka alluded to in his earlier posting above.

I settled on tripoli as the preferred powder, and I presumed it was the same stuff found in liquid metal polish. It was cheap and could be mixed with water, so the PMM was safe from solvents. I added a very small amount of wetting agent to help to disperse the powder in the water, while avoiding a severe dose of the froths.

The mechanism was fairly simple too. I had an old washing machine motor and added a large ball race to the shaft with about 5mm eccentricity. This eccentricity was to produce my up and down movement. A couple of polypropylene (PP) adaptors suitably spigoted to locate into the bore of the valve body closed off the bore, one of which was drilled through to allow the slurry to pass. A simple ball valve sitting inside the bottom adaptor ensured a reasonable one-way flow of slurry. A stubby length of RPVC about 40mm diameter formed the body of the contraption, while providing a miniature priming tank for containing the slurry.

The tricky bit was still to come, ie. how to make the pumping action take place without too much leakage?

As I recall, this came in the form of a couple of rather chunky `O’ rings roughly the same diameter as the RPVC tube and about 6mm in cross section. Besides helping to form a seal between the tank and the outside world, there was sufficient `give’ in the `O’ rings and the rest of the contraption for the eccentric ball race to generate a pumping action. In other words, to squash the assembly up and down, forcing the slurry through the tiny holes.

Fifteen minutes of this while I carried on with something else, and the job was done.

It took a low power microscope to view the beautiful radii which had replaced the burrs and sharp edges on these largely inaccessible places. Although I’ve searched amongst my old samples to provide a picture, it would appear that they have all been dumped during successive moves.

Regards,

Sam

Gentlemen,

In response to a couple of questions about my de-burring technique, especially from Pekka, I cobbled together a couple of schematics from what I remember. It was, after all about 30 years ago.

Michael Williams has reminded me also about the fluidics phenomenon, which I seem to recall involved Enots. However, I hadn’t realised that the valves mentioned could cope with liquids besides air.

Thanks for your comments Michael.

I agree with you about fluidics, and how it doesn’t seem to have established a place in technology. For companies into pneumatics, it was probably attractive in the early days, but I would assume that there are many more electronic switching solutions rather than adopting this method. I've seen (on TV) the results of using air lines to trigger lightening strikes through copper wire, but I imagine even this would not need sophisticated fluidics.

Regards,

Sam