# Relay Output Stepper

#### klangst

Joined Jul 5, 2013
56
New approach to the circuit on the way, better. Still need more information about the test pulse.

ak
Let me know what specific information I can gather for you, if it is in addition to what I just posted.

#### ebeowulf17

Joined Aug 12, 2014
3,275
What are the loads (e-match, etc.) like? What is the typical resistance per load? Do you run multiple loads off a single output? If so, do you use series or parallel connections?

#### AnalogKid

Joined Aug 1, 2013
8,950
What, electrically, are the devices attached to this thing? Are they blasting caps, or strips of nichrome wire that look like resistors, or something else that looks like resistors, or capacitors, or what? Also, clearly 25 mA is not enough to set them off, and someone thinks 3-6A might be needed. Do you have any experience or numbers. Designing an output stage, even one that is a saturated switch, is better when the load is known.

Back to the manufacturer, now that we know what he sends out as a test, the next question is what he needs to see to conclude that the test was successful. In other words, is he looking for some minimum voltage developed across the load to proclaim that a load is connected? If so, what voltage? There are several ways to emulate a load device to keep his system happy, and the more I know about stuff the easier and more reliable it will be.

The schematics are based on a go-to part for some counter circuits because it is counter and decoder all in one. But as you can see, expanding it beyond its natural 10 count limit gets a bit messy. That's why I'm working on a non-4017 version.

ak

#### klangst

Joined Jul 5, 2013
56
From the following site: http://leeds49.co.uk/2013/10/25/ematch-and-talon-igniter-test/

eMatch

12v was applied for 250mS with a max available current of 5A
1. 160mA
2. 152mA
3. 149mA
4. 176mA
5. 169mA
Average ignite current – 161.2mA
Average resistance – 1.1Ω

Talon
12v was applied for 250mS with a max available current of 5A
1. 1.29A
2. 1.56A
3. 1.43A
4. 1.51A
5. 1.37A
Average ignite current – 1.43A
Average resistance – 1.6Ω

Per the control modules specifications, 10 e-match could be connected in series, or 4 talons in parallel.

I typically only use talons and no more than two in parallel.

#### klangst

Joined Jul 5, 2013
56
What, electrically, are the devices attached to this thing? Are they blasting caps, or strips of nichrome wire that look like resistors, or something else that looks like resistors, or capacitors, or what? Also, clearly 25 mA is not enough to set them off, and someone thinks 3-6A might be needed. Do you have any experience or numbers. Designing an output stage, even one that is a saturated switch, is better when the load is known.

Back to the manufacturer, now that we know what he sends out as a test, the next question is what he needs to see to conclude that the test was successful. In other words, is he looking for some minimum voltage developed across the load to proclaim that a load is connected? If so, what voltage? There are several ways to emulate a load device to keep his system happy, and the more I know about stuff the easier and more reliable it will be.

The schematics are based on a go-to part for some counter circuits because it is counter and decoder all in one. But as you can see, expanding it beyond its natural 10 count limit gets a bit messy. That's why I'm working on a non-4017 version.

ak
The talons are simply two wires with a piece of nichrome wire attached between them inside a convenient clip, so they do function as a resistor. I have never measure these myself, but I could get some resistance measurements off of them (I will do that and post the results). The recommended firing current for the talons is 1 to 1.5 Amps. The ematch that I use sometimes (I use talons mostly) have a nominal resistance of 1 +/- 0.2 ohms with a recommended firing current of 1 Amp per the manufacturer (I will check the resistance of some of these as well).

As far as what the controller needs to see, from my understanding, it works like continuity testing and just needs to be a complete circuit. It does not prevent the controller from operating, it will just change the continuity led to red on the controller to alert the operator that the talon or ematch is not showing continuity.

#### klangst

Joined Jul 5, 2013
56
Okay, I checked the resistance of some ematch and some talons. You can see that the talons were not very consistent.

E-match (MJG Firewire Initiator)
1) 1.4 ohms
2) 1.2 ohms
3) 1.4 ohms
4) 1.3 ohms
5) 1.4 ohms
6) 1.4 ohms
7) 1.3 ohms

Talon Igniter
1) 18.7 ohms
2) 10.5 ohms
3) 34.8 ohms
4) 11.3 ohms
5) 4.5 ohms
6) 12.5 ohms
7) 19.5 ohms

#### klangst

Joined Jul 5, 2013
56
Just for the fun of it, I put together what my vision was for the case that will house said circuit. This uses a Seahorse SE-120 case with 8 position under-mount speaker terminals, which I already have a bunch of. This will give me a fairly rugged device that should last me a long time.

#### AnalogKid

Joined Aug 1, 2013
8,950
18 V / 1.2Ω = 15 A. So the devices basically are fuses that are intentionally blown? If so, then the output switch device needs to be able to handle whatever the transient current capability is for the 2 second pulse. Not a problem, just a fatter MOSFET.

A 25 mA current through a 34.8 ohm resistor makes 0.9 V. So to convince the controller that this line has a valid load, we need to sink 25 mA at a compliance of about 1 V. Shouldn't be too hard. Is there some kind of spec or general industry practice that establishes a range for a valid device cold resistance?

Does the operating panel have to be waterproof? Or is the lid open only to make connections, then closed throughout the show?

ak

#### klangst

Joined Jul 5, 2013
56
18 V / 1.2Ω = 15 A. So the devices basically are fuses that are intentionally blown? If so, then the output switch device needs to be able to handle whatever the transient current capability is for the 2 second pulse. Not a problem, just a fatter MOSFET.

A 25 mA current through a 34.8 ohm resistor makes 0.9 V. So to convince the controller that this line has a valid load, we need to sink 25 mA at a compliance of about 1 V. Shouldn't be too hard. Is there some kind of spec or general industry practice that establishes a range for a valid device cold resistance?

Does the operating panel have to be waterproof? Or is the lid open only to make connections, then closed throughout the show?

ak
For what it's worth, most of the controllers I have seen have been designed for a maximum of 10A. This may be due to the ematch with the lower resistance pops almost instantaneously (~250ms or less). See the videos below for a demonstration:

Ematch:

Talons:

I don't know of a industry spec for device cold resistance. If it can "complete" the circuit so that the controller sees continuity, that would at least show that the connection between the controller and the distributor is not disconnected or "broken". The main advantage to the continuity tests is to establish that the ematch or talons show a complete circuit. If they don't, then they need to be checked, adjusted, and in many cases replaced before the show starts. Some devices out there have a separate "continuity" circuit using a 1.5V AA battery to light some LEDs while pressing a momentary switch to show continuity of each output.

While the controller is "waterproof" or water resistant, I don't know that this device needs to go that far. I could have the boards conformal coated and I could "seal" any switches and terminals to help keep moisture out. The SE-120 is a "waterproof" case and I use the SE-300s with all of my controllers and close the lids (but not latch them) once everything is hooked up and verified continuity.

#### AnalogKid

Joined Aug 1, 2013
8,950
Here is the new central core circuit. This is powered from the first controller pulse, and has up to 15 outputs. It needs to be expanded on both ends with indicators and protections, but first lets make sure it does what you need. Note that this circuit switches the loads to GND as we discussed previously.

ak

#### Attachments

• 13 KB Views: 5

#### AnalogKid

Joined Aug 1, 2013
8,950
Notes on the circuit

D3 and R21 prevent the counter from recycling if more than 16 pulses come in. The only way to reset the counter is to have the input stay low until C17 discharges.

R20 and C22 form a 10 ms delay network. This ensures that two output transistors are not on at the same time briefly as the counter changes state.

R1-C1, R2-C2, etc. are lowpass filters to prevent noise from false-triggering an output.

Moving from one output to the next goes like this:

Output 6 is on.
Input pulse goes low while output path to ground still is on.
All outputs are disabled.
Wait.
Input pulse goes high.
Counter increments.
Output 7 turns on.

ak

#### klangst

Joined Jul 5, 2013
56
Nice job ak! I will take some time and go through it in detail and post any questions that I have to help me understand.

#### AnalogKid

Joined Aug 1, 2013
8,950
After D1 I'll add a simple clipper to prevent excessive voltage to the CMOS parts.

ak

#### klangst

Joined Jul 5, 2013
56
I am going to take apart one of the controllers. I am really curious as to which fets it is using as they are able to withstand over 10A with an external power source hooked up.

#### AnalogKid

Joined Aug 1, 2013
8,950
Pictures...!!! The FET part number tells us if it is n- or p-channel, an important detail. Also, while the standard power MOSFET gate voltage range goes to 20 V, most CMOS logic quits at 15 V or 18 V. This raises a subtle question about their control circuits. Pictures...!!!

Don't get too excited about a 10 A part. Power MOSFETs started to take over mainstream switching power supply design in the early 80's. 30 years later, parts with *less than* 0.001 ohm Rdson are cheap:

Toshiba TPHR8504, 40 V, 150 A (continuous), 0.85 milliohm (!!!) - - $2.55, single piece. ak Thread Starter #### klangst Joined Jul 5, 2013 56 Pictures...!!! The FET part number tells us if it is n- or p-channel, an important detail. Also, while the standard power MOSFET gate voltage range goes to 20 V, most CMOS logic quits at 15 V or 18 V. This raises a subtle question about their control circuits. Pictures...!!! Don't get too excited about a 10 A part. Power MOSFETs started to take over mainstream switching power supply design in the early 80's. 30 years later, parts with *less than* 0.001 ohm Rdson are cheap: Toshiba TPHR8504, 40 V, 150 A (continuous), 0.85 milliohm (!!!) - -$2.55, single piece.

ak
I have to get them out of storage (it's about time to anyway). I should be able to post my findings tomorrow afternoon.

#### klangst

Joined Jul 5, 2013
56
I can't make out the fet number on the chips, but one of the fets close has a number of Fairchild FDS4685, which is a p-channel. I did some searching and found FDS4672, which is an n-channel and can take pulse current up to 50A. The 4685 seems to have the same specs but is p-channel. Looking at the outputs, the terminals appear to be switched grounds.

#### AnalogKid

Joined Aug 1, 2013
8,950
I called and talked with Scott. He switches both the high and low sides. The high sides are switched individually with P-channel MOSFETs. I think the low side is a single switch controlled by the ARM key or something like that, but he wasn't exactly clear. He's primarily the software guy, and the circuit was done by "an engineer" who was not available. He doesn't see any problem with us using low side switching, so I'll stick with the n-channel FETs for now.

Even without part-number-clarity, pictures would tell me a lot about the nature of their design.

ak

#### klangst

Joined Jul 5, 2013
56
I called and talked with Scott. He switches both the high and low sides. The high sides are switched individually with P-channel MOSFETs. I think the low side is a single switch controlled by the ARM key or something like that, but he wasn't exactly clear. He's primarily the software guy, and the circuit was done by "an engineer" who was not available. He doesn't see any problem with us using low side switching, so I'll stick with the n-channel FETs for now.

Even without part-number-clarity, pictures would tell me a lot about the nature of their design.

ak
Wow, that's great! Scott is a really good guy to deal with. I sent some pics your way. Let me know if you want anything different.