LED Flasher Relay

[GoldOxide_imp]

Now you confuse people with your wrong voltages and pulse widths. (PB and I are messing with each other here, just so people know).

PB- Tektronix keeps sending me email to buy one of those 10-30 thousand dollar scopes. Like I need one. They even offer to send one to try out. Nope. Mine was about $500 and is good enough for most home stuff. I don’t do any RF at home so 50Mhz is good enough too.

Wouldn’t mind a spectrum analyser but they sure don’t give those away.

There are many low cost spectrum analyzers that are USB-based and connect into a PC laptop computer. Only a few hundred U.S. greenbacks. They are used often for spectral analysis of EMI energy, but can be used in-circuit too. For this CB1100 application, however, it is not needed.

[Lord Popgun]

Yes, Max just recommended one he has. So I ordered one. Much cheaper then the next cheapest thing I could find, a Rigol DSA 705 for ~$700.

[GoldOxide_imp]

Yes, Max just recommended one he has. So I ordered one. Much cheaper then the next cheapest thing I could find, a Rigol DSA 705 for ~$700.

Awesome. And most of those USB-based analyzers can do standard oscilloscope functions too. Example: $500 US might give you a few hundred MHz of bandwidth. Not bad for a mini unit. Even low-end 25 MHz units can do real-time frequency analysis, but obviously bandwidth limited.

In any case, there are so many options on the market these days. Laboratory-grade units must rise well above the rank-n-file to justify their $50K+ prices.

[Lord Popgun]

Tinysa was the one Max has so I ordered one to mess with.

[GoldOxide_imp]

Tinysa was the one Max has so I ordered one to mess with.

Ah ya. A standalone little gem - very low cost. But it does only spectral capture only - right? Nevertheless, sweet tiny unit.

[hardylt_imp]

For those who do not know how to interpret signals on an oscilloscope I’ll try to give a small explanation.

O-scopes display a voltage that takes place over time. The could be constant DC. Could be AC like in your house. Could be waveforms of all types. In our case here, we will see square waves.

The screen is divided into a grid pattern with horizontal and vertical lines. These spaces between the lines are known as divisions. Each of the divisions has ‘tic’ marks. If you start at a line and count tic marks to the next line you will count 5. The value of the divisions, and thus each tic, will depend on scope settings that the user selects.

The trace on a scope can be positioned on any horizontal line. Or anywhere for that matter. Usually the center line is used to put the trace on with no signal, or 0v. And that is what I did for the pics.

When looking at a signal you must select a voltage range. For instance, 10v. Or actually 10v per division. Voltage is displayed going up and/or down. So if you start at one horizontal line, going to the next line is equal to 10v. Each tic would be equal to 2v as 5x2=10. So 12v will show as one division plus 1 tic. Other voltage levels can be selected.

Time goes from left to right. You select a time base that will show the signal as you want. So, for instance, 5ms (milliseconds) So from one vertical line to the next would take 5ms. Each tic would be 1ms. A pulse that is 1 division and one tic would be 6ms.

In the pics the pulse trains look pretty much the same until you see the time base is different, thus the pulse widths and repetition rates are different. I will point that out with the pics in a following post as I don’t know how long a post can be.
Test equipment: Tektronix TBS 1052 dual trace 50MHz digital scope.

Test setup: Scope probe connected to the left turn signal orange wire in headlight shell. Probe ground lead connected to the ground plate in the left turn signal (easy to get to and the lens was already off).

Bike-battery tender disconnected. Ignition key on, bike not running.

Scope set to 10v per div and 5ms per div.

[url=https://app.photobucket.com/u/everettmclendon/a/a8b5b03b-ad65-49a1-a3d0-f45df1fc2c20/p/75178ebe-275d-493f-aea0-37df9fe25eff]

First, The settings shown should be /division not /cm I had centimeters on the brain last night.

In the pic the running lights are on. The picture shows that they are light by a pulsed 12vdc. The pulses show the voltage is on for 2ms and off for 4ms. Since the voltage is not a constant 12v they are dimmer than they would if fed a constant 12v.

[url=https://app.photobucket.com/u/everettmclendon/a/a8b5b03b-ad65-49a1-a3d0-f45df1fc2c20/p/36df8cf5-894b-45b6-9822-c698cc4dee01]

The above pic is with the left turn signal selected.
The scope is set to 10v/div and 500ms/div

The turn signal light flashes by the voltage turning on and off (obviously) It is on for ~400ms and off for ~300ms or on about .4sec and off for .3 sec.

You can see the pulses are 12v as they are 1 division high plus 1 tic. That is 10+2=12

Any mistakes blame PB

Fascinating! Thank you for that investigation LPG. Couldn't have asked for a more in-depth explanation. And I appreciated the crash course in oscilloscopes as well.

The flicker makes more sense now. 2ms is nowhere near enough time for the sweeping pattern on my LED's to complete. And, the whole LED array doesn't flicker--only the first diode in the pattern. So if I understand it correctly, every 6ms or so it's just constantly trying to begin that flash pattern, with never enough time to complete. And so instead I get a rapid flicker on the inside corner of both signals. Mystery solved.

I imagine if it was a more typical LED signal with a uniform flash pattern, not sweeping, the running light power would possibly translate into a more useable low-level illumination. Or at least all diodes would flicker together? But overall I get the sense that 2-wire LED's are probably just not intended to function as both DRL's and turn signals in that way. Seems like that's what 3-wire LED's are for.

[Lord Popgun]

Yeah, that was my thought as well Hardy. I think just wiring them into the rear signals, as Max suggested is going to be your best/easiest way.

[hardylt_imp]

Yeah, that was my thought as well Hardy. I think just wiring them into the rear signals, as Max suggested is going to be your best/easiest way.

Worked like a charm. Thanks all. To anyone in my position in the future--POSH relay #090015-W1 is what I tried as a solution, and might be the easiest fix to any flasher relay problems--pricy little thing, but it's also the only LED relay out there that fits our stock connector pins as far as I can tell after much searching. It will save you installing resistors on the front two signals that get very hot very fast from running light power. Good luck with the fuel tank quick disconnect though. It was a real doozie for me, disconnecting and reconnecting.

[Run Away_imp]

Interesting thread.

I just stumbled across it while looking up info for converting my turn signals.

I have a 2018 CB1100RS, Canadian market. It has the older style incandescent orange coloured signals. I bought a set of used OEM Honda turn signals from a Japanese-market CB1100RS, which are LED from factory.

Here's a picture of the two types next to each other.


Notice the JDM signals have an extra 3rd wire on the front signals, possibly for the DRL function? My bike doesn't have any blank plugs hanging out behind the headlight that match. The pulse width modulated running lights in an interesting discovery that explains some of the things I've found.

I searched an online Honda dealership in the UK and looked up the part number for the UK spec turn signal flasher for a 2018 CB1100RS. The part number is 38301-KZZ-J01, and the catalogue description says it's for LED signals. It seems it's used on a few Honda bikes.
Here's a pic grabbed from Google


So my theory is that on CB1100s that come with factory LED turn signals, they do away with the pulse width modulated DRL function, and add that feature back in with a 3rd wire to the front lights. This allows them to use a more conventional flasher relay shared on other bikes, which has fewer wires.

[hardylt_imp]

Interesting thread.

I just stumbled across it while looking up info for converting my turn signals.

I have a 2018 CB1100RS, Canadian market. It has the older style incandescent orange coloured signals. I bought a set of used OEM Honda turn signals from a Japanese-market CB1100RS, which are LED from factory.

Here's a picture of the two types next to each other.


Notice the JDM signals have an extra 3rd wire on the front signals, possibly for the DRL function? My bike doesn't have any blank plugs hanging out behind the headlight that match. The pulse width modulated running lights in an interesting discovery that explains some of the things I've found.

I searched an online Honda dealership in the UK and looked up the part number for the UK spec turn signal flasher for a 2018 CB1100RS. The part number is 38301-KZZ-J01, and the catalogue description says it's for LED signals. It seems it's used on a few Honda bikes.
Here's a pic grabbed from Google


So my theory is that on CB1100s that come with factory LED turn signals, they do away with the pulse width modulated DRL function, and add that feature back in with a 3rd wire to the front lights. This allows them to use a more conventional flasher relay shared on other bikes, which has fewer wires.

Makes sense. Thanks for sharing. I was unaware that some of them come stock with 3-wire LED’s, sure would have made my project easier. Well I hope that flasher you found works with the Japanese LED’s you got, I’m curious to know if they do