Site Map FAQ
E-mail Ray

Valid XHTML 1.0!


Updates on the aRocket discussion list group project.

Andrew Case has put a page up on his igniter progress

Duncan McDonald has a section at the Sourceforge.net RocketWork Bench site and is blogging his progress there.

The aRocket group has begun work on a reusable open source ignitor design. Here are the updates posted to the aRocket listserv:

Posted June 28, 2003 by Duncan McDonald

Did some more testing of the coil/sparkplug yesterday. Put some thermal goop on the heatsink which kept the mosfet cooler. Just used the Ford "coil on plug" as it has the best performance of all the coils tested. It is interesting that all the other coils had turns ratios of about 50 whereas the Ford unit had a turns ratio of almost 70. That combined with a 38mH magnetizing inductance makes this one a winner. Now I just have to find the Ford part number and how much it costs. Played around with the two best spark plugs from last time, the BM6F and the 10mm NGK racing plug. Somehow the racing plug didn't seem quite as impressive as it did the first time we tried it. Still produced a very hot spark, but it was more or less the same as the BM6F. The racing plug has a drawback of a 3/4" threaded barrel ("reach" in sparkplug nomenclature) but not having a protruding electrode should help avoid any potential fouling issues. As a result we decided to go with the NGK plug. It's less than $3 so it is certainly cost effective. Tried various duty-cycles and found a range of 40 to 60% would work. 60% gave a bit hotter spark, 40% a bit weaker. At 10% it was extinguished or close to it. 1KHz was the best overall frequency so we decided to just go with 1KHz and 50% duty cycle.

Worked through the preliminary design of the controller board. It will have something for everybody (at least the prototype): AVR controller, CAN bus interface or parallel control lines, two channel temperature sensing (chamber and heatsink), two channel solenoid control, status readback via the AVR or parallel lines. Should have boards in two weeks. Question though: are there strong preferences for SMD or through-hole? Either will work and through-hole may have some advantage because it is easier to heatsink through-hole devices. But it will be smaller with SMD. Any opinions?

Thom has a design for the chamber and is starting on that. It'll be GOX and Hydrogen. The prototype will have outboard gas regulation so we can easily change pressure. The body will be a chunk of stainless steel with diffusion welded SS tubing to the gas ports. Holley solenoids for gas control. Looking at reusing large threaded CO2 cartridges as the flight gas storage. Will start testing the prototype chamber a week from Monday (July 7).

So far so good!


Posted June 24, 2003 by Dr. Case

I just got the regulator I ordered from Aerocon. Looks good, if a bit larger than I expected. It looks like cleaning it for Oxygen service won't be too hard, though I haven't taken it apart yet. That will have to wait until I can get it to a cleaner environment - no point in stripping it until I'm ready to clean and put it straight into a clean, dry ziploc bag.

It needs gauges and fittings, which I bought from McMaster-Carr. Interestingly enough, despite the fact that it's easy to pay upwards of a hundred bucks for gauges, McMaster offers Oxygen cleaned high pressure gauges for under $12 a pop - I bought two, a 4000 psi one (part # 32255K77) for the high pressure side and a 1000 psi (part # 32255K84) for the outlet. The gauges are good to only 3% midscale (which means probably at least 5% at the ends of the scale), but 3% is good enough.

Posted June 16, 2003 by Dr. Case.

I've pretty much finished the igniter body and injectors. The only work that still needs to be done is a little tidying of a burr on the Isopropyl injector and fitting a plug to the ignition detection port. Once I've fired successfully I'll start work on ignition detection, but for now I'm just going to blank off the port.

There are a couple of imperfections in the injectors and body, all of them due to thinking while standing in front of the milling machine. Never do this. Think at the desk, machine at the mill. I relearn this lesson about once every ten months or so. Redesign on the fly as you are making a part is just not smart. As it turns out the imperfections have no functional implications, but it's irritating to put so much effort into planning only to make simple errors.

I've ordered a regulator for the GOX, Aqua Environment model 415-1500 from Aerocon Systems. Price was $260 including S&H. It's not O2 cleaned, but I can do that myself. Also ordered a Lake Axial Flow solenoid, which is actually for another project but might get used on the igniter depending on how things work out.

I posted last week about the conical end mills from Conical Tool. They've arrived, and they look like just the ticket for the job. The smallest has a tip diameter of 1/32" - very small and delicate looking.

Once the regulator arrives the remaining items before test firing are: machine a mount for the igniter find a suitable tank for the IPA design and build a mount for the IPA tank decide on IPA pressurization system and implement it hook up plumbing sort out electronics

The last item may turn out to be trivial if I just use the preexisting circuit from my previous effort. This may be the best approach, since the using the AVR board requires some on-board construction and learning to program in either AVR assembler or c. Once I'm confident of the other parts I can tackle the microcontroller and associated circuitry. IPA pressurization is likely to be via something cobbled together from paintball gun parts.

Date: Thu, 29 May 2003

From: Duncan McDonald

Can't say I got a whole lot done last week. This week is heading in the same direction. I did get together with Thom McGaffey and remeasured the coils DC resistances. Thom has a 6 1/2 digit voltmeter and figured out how to set up 100 power line cycle integration for stable low resistance measurements. The new measurements confirm the previous set: on an automotive coil the primary + terminal is the common connection for the HV secondary. I received my parts from Digikey so I will try and build a breadboard of the coil driver this week. Like Andrew, I have had my eye on the CM6 10mm sparkplug. Bill Bullock came up with a fantastic website http://www.sparkplugs.com/. Click on "Advanced Tuners and Engine Builders Data" and there is complete parametric selection. Check out the 10mm racing-style plugs for $2.29 http://www.sparkplugs.com/more_info.asp?AAIA=&pid=3315.

Date: Wed, 28 May 2003

From: Andrew Case

I was ill all of memorial day weekend, so the expected burst of productivity never materialized. What I did get done was finalizing the injector design and locating some sources of information on GOX handling. I just received a copy of Vance Harlow's Oxygen Hacker's Companion, which is actually for divers, but contains lots of useful information, particularly for people planning to reuse diving gear in nontraditional ways :-)

I received the spark plugs I ordered, NGK CM-6 10mm spark plugs. They are really small, so I'm glad I didn't go for the 8mm size. These are on the borderline of getting finnicky to handle. One unfortunate discovery is that the threads are 10mm by 1mm pitch, so the hole I tapped 10X1.5 is going to be drilled out and tapped 1/4 NPT for the ignition detection, and a new hole drilled and tapped 10X1 for the spark plug.

This week's plan is to build the injector head and start familiarization with the AVR.

Date: Tue, 20 May 2003

From: Andrew Case

I'd designed the new igniter body, and built most of it. Still to do are a tapped port for the ignition detection and four tapped holes to hold on the head end. I've got numbers and a rough design for the injectors, nominally using GOX/IPA. One immediate insight from the injector design is that the holes are really small, and will be a major PITA to machine. I'm working on refinements of the head end and injectors to set things up so that the tricky machining is spread out over multiple pieces, and there is no tricky stuff that has to be done right at the end of a whole bunch of machining. This is to try to minimize the probability of investing huge amounts of time in the piece only to blow it on some finicky but critical detail. If I'm going to mess up I want to do it early, so I can start over with minimal waste of time. Also I think I have found a way to finesse things so that I can cut and try on the Oxidizer injector.

I'm going to double check my calculations tonight and I'll post calculations and dimensions tomorrow unless something comes up.

The AVR board, programmer, and chip are now in my hands, and waiting for me to have some time to work on learning a little about microcontrollers and programming them. Jamie Morken has sent me some sample code to look at. If things go smoothly with the injector design and construction (ha!) I'll start looking at the microcontrollers this weekend. I'm guessing the injectors are going to be a bit of a production to do, though. I also have Mike Castle's PIC board waiting for me to take a look at it. I'll do that in parallel with the AVR.

Date: Tue, 19 May 2003

From: Duncan McDonald

(You'll have to go over to http://rocketworkbench.sourceforge.net/projects.phtml to get the files referenced below. The reports will be blogged there as well.)

I went over to a company owned by some friends of mine (Amicronix Test Systems) that has a small metrology lab. They have an HP 3458A 8 digit voltmeter that will resolve 10 micro ohms. I hadn't received the mosfets from Digikey so I couldn't take the inductance meaasurements. I settled for taking the equivalent series resistance (ESL) measurements of the coils primaries and secondaries and measuring the coil turns ratios. I first measured the turns ratios by driving the primary with a small (40mV to 250mV) sine wave and measuring the secondary peak-to-peak voltage. There was a lot of "smear" on the low voltage waveform that made the measurement difficult. I then tried a higher voltage of 1Vp-p on the primary and measured the secondary again. These "high drive" measurements came out much better. I recorded the voltages and the respective phases in the attached spreadsheet. The phase gives the "dotted" sense of the terminals. Transformeers are notated with dots on the windings to show the relative phase of one end of each winding. As I expected, the secondary (high voltage lead) was in phase with the "+" side of the primary on the two coils that had a marked polarity.

Next, I measured the primary and secondary ESR's using a four-wire kelvin connection. I was able to decently resolve 1 milli ohm, but below that the measurements were a little noisy. The HP 4184 has the provision to increase the ADC integration time, but I couldn't figure out the menu system. I also couldn't find the manual (almost everyone had already left for the day). On a whim, I decided to measure the secondary with respect to both ends of the primary (+ and -). The secondary impedance was always higher with respect to the "minus" side of the primary than the "plus" side! I was surprised, because that means the + side is the common point with the secondary, just the opposite of what I would have thought (see attached schematic). To read the schematic you'll have to download the viewer from ExpressPCB. I seem to recall that automotive sparkplugs fire when the electrode is negative so electrons flow from the center electrode to the rim. Is that correct? If so, this picture is consistant with that scenario. But also if this is true, I should really redo the turns ratio measurements as I may have been driving the coil in such a way as to bias the secondary with the primaries drive level. It's a small error, but should be corrected.

I am not 100% sure of my results because of the noise problem I had with the voltmeter. Because of the noise I really only had 4 significant digits. I need at least one more to really be sure of these results. To reduce noise I either have to average some samples or increase the integration time (the voltmeter has a dual slope integration converter). I couldn't figure out how to do that by just fiddeling around, I need to read the manual. Unfortunately Amicronix just sent the meter out for calibration so I can't use it again for a couple of weeks. Thom has a 6 and a half digit voltmeter so I am going to go over to his house on Wednesday and try again. If the IGBT's come in we can also do some inductance testing. Thom also reports he received the thermocouples for ignition detection.

I would also like to get some pictures posted this week!

Date: Tue, 6 May 2003

From: Duncan McDonald

I came across this technique for non-optical ignition detection that apparently is used in "smart engines". During the portion of the ignition cycle when there is no spark, ie when the coil is charging, place 400V across the plug spark gap and see if about 1mA of current flows due to gas ionization. That does require some pretty heavy duty diodes to keep the HV out of the 400V supply. But how about having a second spark plug that just does ignition detection? There is still the matter of a 400V supply, but an LT3420 would do nicely.

See http://www.fs.isy.liu.se/~larer/Projects/main.html And http://www.hut.fi/~vvartiov/ion/DIY-Ion-Sensing-2.pdf. Clever use of a current mirror for current sensing. Note that the current mirror is drawn incorrectly. The BF421's are PNP transistors.

Seems a little complicated, but in the long run maybe it is simpler than trying to create an optical window into the combustion chamber. What do you guys think?

Date: Tue, 5 May 2003

From: Andrew Case

I just uploaded some PIC code to the arocket uploads area. This code was written by Mike Castle to sequence valves and run the spark using a PIC microcontroller. The file is MCPICign.txt.

I haven't run it, but I will when I get set up to do PIC stuff. I'll report results to the list.

Date: Mon, 4 May 2003

From: Andrew Case

First the controller: I've ordered the programming cable and prototyping board (available through the Olimex distributor in the US at www.sparkfun.com) along with an ATmega8 chip to populate the board. Should be arriving sometime this week. Also Mike Castle in the UK has a PIC board that he's offered to send me (along with some extras for distribution), so we can have a shootout :-) In fact I think I'll probably settle on the AVR, but if we can get a PIC variant running, the more the merrier - the fact that there are some extra free boards lowers the entry barrier for people who want to build whatever we come up with (this is important since the more people building hardware the more thorough the debugging will be).

Ignition detection: I wasn't able to find a photodiode since I couldn't get to my regular electronics supplier before they closed, so I went to radio shack and picked up a phototransistor and a photoresistor. I set them up one at a time on my existing igniter, with the propellants off - just testing if I can see a spark. The bad news is that the electrical pickup from the spark line dominates the signal on both. This isn't all that surprising since I had big loops of wire everywhere. Obviously optical detection will require a bit of finesse, if it works at all. I'll try again down the line, probably with a photodiode and better shielding. It's going to require a bit of effort, and it's not on the critical path, so I'm not going to put a really high priority on it.

Date: Sun, 3 May 2003

From: Andrew Case

I just toasted another 555 due to lousy fit of a component carrier in a socket. I'm considering replacing the sockets, but perhaps it's time to go to a better system. The problem I have with the various microcontroller suggestions so far is price. All the ones I've looked at closely have expensive programming hardware. That need not be a huge obstacle since there are presumably people willing to rent out the use of their hardware, but that takes time. I need to be back on the air by next weekend at the latest, so I need a COTS solution.

I found a possibile solution in the Basic Stamp: . It's only $79 for the bottom end of the line, including programming software, manual, and cable. Memory holds only 80 instructions, but the application is simple. There's also a little space for interface electronics. One nice thing is that it runs directly off a 9 Volt battery, though power can be 6-15 Volts. If you torch the chip, replacements are only $34. The chip executes ~2000 instructions per second, and it loops continuously through the code in memory, so even if you use the full 80 instructions you're looking at executing the whole loop every 40 milliseconds, which is fast enough. In practice you only need to check the status of the ON switch and the ignition detector, and sequence the valves in the right order, so you ought to be using only 40 instructions.

Questions: Is there a catch somewhere with using the Basic Stamp? Are there other cheap, simple, ready-to-go microcontrollers?

Date: Thu, 1 May 2003

From: John Carmack

XCOR uses a programmable industrial sequencer to run their igniter, but I, of course, wrote a computer program to control it from a laptop.

I use the parallel port for input and actuation (I use a little public domain driver to allow this under Win2K / WinXP, like it used to be allowed in Win9X).

bit 0: igniter gox bit 1: igniter fuel bit 2: main propellant oxidizer bit 3: main propellant fuel bit 4: igniter spark (the computer toggles this, so it doesn't need any extra timer circuitry) paper out input signal: igniter chamber pressure switch

Our current setup is slightly different, because we have hacked up one of our old vehicle driver boards to do this, but we are probably going to build a nice new board that includes two bi-directional motor drives so we can control our big servo valves with the same board. Russ is also looking into building the spark generation hardware directly, instead of using the MSD equipment. Russ's company makes some stun gun type hardware as well as radio hardware.

I am going to integrate support for dataq serial logging this weekend, so it will be a nice, self-contained little program like our old test stand program.

I will make this source code public domain for the igniter project after we get everything tested and cleaned up a bit. The drive board for this will probably cost a few hundred dollars for all the solid state relays and such, but we can probably offer the layout and parts list to the project.

Date: Thu, 1 May 2003

From: Andrew Case

I got a quote from Acra-Ball on glass balls at less than 2 bucks each - very reasonable, except the minimum order is $100. I'm trying other options for a viewport.

Pierce reminded me that XCOR used Ethane for their teacart engine - self pressurizing to ~540 psi, nontoxic, pretty nice stuff. Why don't more people use it? - Roberts Oxygen (a local gas supplier) just quoted me ~$200 for a three pound bottle. That's why :-(

Still, it might be worth it for debugging and perhaps for use in small quantities such as for lighting hybrids. I'll talk to SWMBO about financial priorities. Maybe it's time to invest in a regulator and a pressurizing system and just go straight to Isopropanol.

Date: Thu, 1 May 2003

From: Andrew Case

For people working on electronics boards who might be interested in having me test them on my system, here are the magic numbers:

I currently switch my solenoids with 7.4 Volts from a rechargeable battery, so the switching pulses to the transistors need to be at least 7.4 V. My present controller uses 9V and that works fine. The timing sequence is: Spark train starts simultaneous with opening the oxidizer valve, followed 0.1 seconds later by opening the fuel valve. Spark frequency is 500 Hz.

Ignition detection is a complicated issue, best handled separately. I can provide a 5V signal that goes high when ignition is detected.

Right now I can test with either a full HV system, or with something that just gives a pulse train out. Since I have an MSD ignition I'll just pass the pulse train to the ignition module (which is what I'm doing now).

If need be I can cobble together something to translate 5V pulses into voltages which are more useful to me.

Date: Wed, 30 Apr 2003

From: Duncan McDonald

I'm about ready to start doing some work on the high voltage portion of the igniter and I'm thinking about two versions: a dumb version and a smart one. They could probably be the same PCB, just more parts stuffed in the smart one. The dumb version has a two pin input connector so that when you apply a logic one (+5V), the igniter starts producing HV. In the smart one there is a comm link and processor that receives a command like "$ignition1 on". The processor then starts the HV. The smart version could have some options like measure HV current flow and HV voltage produced, igniter gas valve sequencing, etc.

-Any particular coil that is small and inexpensive? The Fiero one mentioned is $45- 50. -Seems like 0.030 is a pretty common plug gap which translates to about 17kV, so the coil output should be 20-25kV min. -9V to 12V input voltage to the board. -For the smart version, what should the comm protocol be? TCP/IP or UDP/IP? That seems kind of complicated, but the additional stack code bumps up the processor requirement to an Atmega123 ($17, based on the stack taking 8 to 16k memory) plus an interface IC like the Cirrus 8900A. So that is maybe $13-$16 more than a bare bones serial link. Not unreasonable. I believe a CAN bus implementation would be a bit less $$ as a less hefty processor is needed but you still need one or two interface IC's.

Thoughts? Other ideas?