Fuel injector test pump

[Roger B]

There are several people on here interested in miniature diesel fuel injection systems. One of the first requirements for this is a test pump that will reach at least ~80 bar (1200psi). This is derived from the compression pressure required to cause ignition in a small engine of at least 30 bar (450psi). The drawings are below in .DWG and .PDF. This is based on information I have already posted on MEM.

Most of the dimensions in this drawing are not critical. It is useful if the operating lever has a simple ratio and low friction so that the force required to inject can be measured with a spring balance or hanging weights. I used a 6 x 2 x 3mm miniature ball race to push on the pump piston, this could also be a hardened steel disc. The operating arm has a bronze bush for the pivot.

The frame is a short piece of 50mm (2”) angle iron. The body is 10mm square hot rolled mild steel bar. It could be 12mm or ½”.

The operating arm is cut from 6mm aluminium. The pivot pin, tappet, excentric stop and the pin for the tappet bearing are silver steel (drill rod), mild steel would probably work for all these parts.

The inlet and delivery connections are made from modified commercial M5 x 0.5 brass pipe unions.

The delivery valve spring is:
Wire diameter 0.4mm
OD 2.9mm
ID 2.1mm
Free length 6.3mm

The return spring is:
Wire diameter 0.8mm
OD 8.6mm
ID 6.9mm
Free length 13.5mm


The small excentric stop behind the operating lever allows the piston to return so the inlet port is fully open.

I used Acro needle laps and Bal-tec ball laps:

https://acrolaps.com/index.htm

https://www.precisionballs.com/ball_valve.php

I would suggest starting the body from the delivery valve end, it can then be turned round to machine the spring seat as the concentricity is not so important.

 

[Roger B]

For the original pump I used 2mm silver steel (drill rod) for the piston, hardened and then lapped to 1.98mm with a simple split lap held in a die holder. I then had a problem with a piece of trilobal (a problem due to centreless grinding where it measures correctly with a micrometer but is not round) bar and after that I used commercial pin gauges for the pistons. These can be cut in the lathe using an abrasive disc in a Proxxon/Dremel type hand tool. The end to be threaded M2 can be annealed in flame, whilst holding the rest in some heavy pliers to act as a heat sink. The tappet is fixed flush to the piston with a little Loctite on the M2 thread. I lap the piston after the cylinder is lapped so I can check the fit. If a pin gauge is used there is no need to lap the piston.

The cylinder bore is the most complex part, the drilling, reaming and lapping is very much a hand skill. I made many test pieces to understand how to drill and ream to get a true, parallel bore. A set of pin gauges in the appropriate size range, in this case from 1.95mm to 1.99mm is essential. This was my best sequence:

Run in the lathe at 2000rpm.

Centre drill with a 1.2mm tip bit deep enough to leave a small countersink.

Drill 1.85mm.

Ream 1.95mm with a controlled feed of 0.08mm per rev (set the saddle to this feed and follow it with the tailstock).

This seemed to work, the 1.96mm pin gauge passed straight through, the 1.97 mm barely entered.

A floating reamer holder helps to avoid bell mouthing and an oversize hole if the headstock and tailstock are not well aligned but these are hard to find in small sizes and expensive.

I made a range of tests with different size drills and techniques to reach my optimum, you may have to repeat this with your lathe.

The next step is to drill the 4.5 mm hole for the delivery valve, flatten the bottom of the with a 4mm, or better 4.5mm, end mill and tap M5 x 0.5. This sequence removes the possibly slightly bell mouthed start to the cylinder.

The inlet is next, drilled and tapped M5 x 0.5, and then the 1mm port is drilled through into the cylinder. Run the reamer through again by hand to remove any burrs.

The bore is then lapped using an Acro needle lap and diamond paste, starting with 5 micron and finishing with 1 micron. I hold the lap in a small drill chuck in the headstock and run the lathe at 500rpm. Put something over the lathe bed to protect it from drips of splashes of abrasive. Checking the progress of the bore is difficult as measuring tools like pin gauges should not come into contact with abrasives but I think as long as everything is cleaned up afterwards in an ultrasonic bath it is a limited problem. Keep lapping until the 1.98mm gauge just passes through and the 1.99mm gauge won’t enter. A few drops of light oil will be needed from time to time.

Finally drill and tap the M3 fixing holes, taking care not to break through into the cylinder. Turn the spring seating on the other end of the cylinder. These operations could be done earlier in the process.

The delivery valve seating can be made by putting a ball in place and tapping it squarely with a brass drift, then throw that ball away. Alternatively, a ball lap can be used.

The thread on the inlet union is shortened to clear the unfinished thread at the bottom of the port. The delivery valve/union is drilled out to 3.1mm for the ball and spring and the thread is machined away at the cylinder end to clear the unfinished threads and make a narrow sealing anulus.

Clean all the components in an ultrasonic bath to remove all traces of the diamond paste.

As a test fill the pump with a light machine oil, get all the air out and fit a blank to the delivery union. It should support a load of 2 or 3kg on the plunger for at least a minute. If it sinks more rapidly the piston/cylinder fit is not good enough.

The pump in use:

 

[The Ignoble Troll]

Thank you for going through the effort of making such a thoughtful post. Greatly appreciated!

 

[ruben]

For the original pump I used 2mm silver steel (drill rod) for the piston, hardened and then lapped to 1.98mm with a simple split lap held in a die holder. I then had a problem with a piece of trilobal (a problem due to centreless grinding where it measures correctly with a micrometer but is not round) bar and after that I used commercial pin gauges for the pistons. These can be cut in the lathe using an abrasive disc in a Proxxon/Dremel type hand tool. The end to be threaded M2 can be annealed in flame, whilst holding the rest in some heavy pliers to act as a heat sink. The tappet is fixed flush to the piston with a little Loctite on the M2 thread. I lap the piston after the cylinder is lapped so I can check the fit. If a pin gauge is used there is no need to lap the piston.

The cylinder bore is the most complex part, the drilling, reaming and lapping is very much a hand skill. I made many test pieces to understand how to drill and ream to get a true, parallel bore. A set of pin gauges in the appropriate size range, in this case from 1.95mm to 1.99mm is essential. This was my best sequence:

Run in the lathe at 2000rpm.

Centre drill with a 1.2mm tip bit deep enough to leave a small countersink.

Drill 1.85mm.

Ream 1.95mm with a controlled feed of 0.08mm per rev (set the saddle to this feed and follow it with the tailstock).

This seemed to work, the 1.96mm pin gauge passed straight through, the 1.97 mm barely entered.

A floating reamer holder helps to avoid bell mouthing and an oversize hole if the headstock and tailstock are not well aligned but these are hard to find in small sizes and expensive.

I made a range of tests with different size drills and techniques to reach my optimum, you may have to repeat this with your lathe.

The next step is to drill the 4.5 mm hole for the delivery valve, flatten the bottom of the with a 4mm, or better 4.5mm, end mill and tap M5 x 0.5. This sequence removes the possibly slightly bell mouthed start to the cylinder.

The inlet is next, drilled and tapped M5 x 0.5, and then the 1mm port is drilled through into the cylinder. Run the reamer through again by hand to remove any burrs.

The bore is then lapped using an Acro needle lap and diamond paste, starting with 5 micron and finishing with 1 micron. I hold the lap in a small drill chuck in the headstock and run the lathe at 500rpm. Put something over the lathe bed to protect it from drips of splashes of abrasive. Checking the progress of the bore is difficult as measuring tools like pin gauges should not come into contact with abrasives but I think as long as everything is cleaned up afterwards in an ultrasonic bath it is a limited problem. Keep lapping until the 1.98mm gauge just passes through and the 1.99mm gauge won’t enter. A few drops of light oil will be needed from time to time.

Finally drill and tap the M3 fixing holes, taking care not to break through into the cylinder. Turn the spring seating on the other end of the cylinder. These operations could be done earlier in the process.

The delivery valve seating can be made by putting a ball in place and tapping it squarely with a brass drift, then throw that ball away. Alternatively, a ball lap can be used.

The thread on the inlet union is shortened to clear the unfinished thread at the bottom of the port. The delivery valve/union is drilled out to 3.1mm for the ball and spring and the thread is machined away at the cylinder end to clear the unfinished threads and make a narrow sealing anulus.

Clean all the components in an ultrasonic bath to remove all traces of the diamond paste.

As a test fill the pump with a light machine oil, get all the air out and fit a blank to the delivery union. It should support a load of 2 or 3kg on the plunger for at least a minute. If it sinks more rapidly the piston/cylinder fit is not good enough.

The pump in use:

Thank you very much for these plans and the explanation of how to do it. few people publish it thank you