Cluster inverter reverse engineering project

mwolson

Gray beard
Established Member
Joined
Mar 28, 2004
Messages
3,411
Location
San Jose, CA
I have decided to reverse engineer the inverter for the electroluminescent backlight in the Termi instument clusters. I have a working inverter that I can use to see what the signals should look like, but I need a couple of dead units to I can take them apart to see how the transformer is wired and to try to figure out what has failed.

Assuming I am successful, I will publish the results here so those who follow will be able to fix their inverters when they fail in the future.

If anyone has a dead inverter they can spare, please PM me so we can make arrangements to get it to me.

Thanks in advance.
 

mwolson

Gray beard
Established Member
Joined
Mar 28, 2004
Messages
3,411
Location
San Jose, CA
My customer's cluster had a bad EL panel it turned out. I swapped his PCB into a junkyard donor cluster to get him back on the road. I have learned that you should measure many MOhms of resistance across a good EL panel. I read about 14KOhms across the bad panel, so low resistance is a symptom of a bad EL panel, not a bad inverter.

It isn't always the inverter that fails. With the good EL panel, my DVM said it was about 43MOhms which then went up to about 100MOhms and then went off the scale.

My customer had assumed that the inverter had failed, so he had bought a SpeedHut inverter, which he included when he shipped the cluster to me. It turned out that the new SpeedHut inverter was also dead so I couldn't see what kind of signal it puts out. SpeedHut has sent a replacement inverter to the customer.

I was corresponding with SpeedHut support about this issue. People have noticed that their clusters are dimmer with the SpeedHut inverter. The SpeedHut support person said that the EL panels brighten and dim based on the amplitude of the sine wave going into the panel. SpeedHut says that the amplitude of the output of their inverters should be 100-200 VAC.

I was able to get a scope on AC signal going into the working cluster EL panel with a working Ford inverter. I saw a 300VAC P-P, 450Hz signal on the scope:

1658276879342.jpeg


When I unhooked the EL panel to see what the Ford inverter puts out with no load, it shot up to over 400VAC, which is the limit that my scope can see:

1658276990782.jpeg


While it is safe to check the resistance of the EL panel with an ohmeter, it is NOT SAFE to work on a powered up inverter. 100VAC can kill you and >400VAC can really kill you. So BE VERY CAREFUL if you decide to test an inverter yourself. Make sure you are not alone working on this stuff and make sure that whoever is with you is CPR trained. I lost a fellow grad student in grad school who was working on a high power laser power supply. We all learned CPR after that. I am not responsible if you get yourself killed working on your instrument cluster.

Also, make sure that your test equipment can handle at least 1000VAC if you are going to test one of these yourself.

Just a reminder, I am still looking for any dead inverters I can get so I can dig deeper into reverse engineering these things.
 
Last edited:

imacgyver

New Member
Joined
Jun 6, 2022
Messages
3
Location
Bermuda Dunes, California
I’m the one who owns the ‘04 SVT Cobra that Mark is describing above. Just wanted to let all of you know what a pleasure experience it was to work with Mark. I sent him my Cluster and SpeedHut Inverter, and less than 5 minutes after I received a delivery notice from USPS he called me. He had a lot of ideas, and was eager to put them into motion. I received the Cluster than he created for me from two today, and it is already in and working...all at a very reasonable price. Easily the best repair experience I have ever experienced!
 

shurur

Well-Known Member
Established Member
Joined
Jan 2, 2011
Messages
3,760
Location
Lutz, FL
I’m the one who owns the ‘04 SVT Cobra that Mark is describing above. Just wanted to let all of you know what a pleasure experience it was to work with Mark. I sent him my Cluster and SpeedHut Inverter, and less than 5 minutes after I received a delivery notice from USPS he called me. He had a lot of ideas, and was eager to put them into motion. I received the Cluster than he created for me from two today, and it is already in and working...all at a very reasonable price. Easily the best repair experience I have ever experienced!
You are welcome..just kidding... ;-)
 
Last edited:

98 svt

Well-Known Member
Established Member
Premium Member
Joined
May 16, 2005
Messages
23,964
Location
Massachoooosetts
I’m the one who owns the ‘04 SVT Cobra that Mark is describing above. Just wanted to let all of you know what a pleasure experience it was to work with Mark. I sent him my Cluster and SpeedHut Inverter, and less than 5 minutes after I received a delivery notice from USPS he called me. He had a lot of ideas, and was eager to put them into motion. I received the Cluster than he created for me from two today, and it is already in and working...all at a very reasonable price. Easily the best repair experience I have ever experienced!


Mark is a great guy. He has been a staple of SVTP for many years.
 

mwolson

Gray beard
Established Member
Joined
Mar 28, 2004
Messages
3,411
Location
San Jose, CA
In the SVTP thread that tells how to use a SpeedHut inverter to replace an OEM Ford inverter, people have said that their clusters are dimmer after the replacement.

imacgyver sent me the replacement SpeedHut inverter yesterday and this one does work. I had also bought an EL panel from a company called GlowHut so I would have a panel I can use for experiments, so I put the stock OEM Ford inverter and the SpeedHut inverter on it so I could compare the results to see if I could explain why the SpeedHut inverter is dimmer. Here are the results:

This photo shows the Ford OEM inverter driving the GlowHut panel:
1659141659020.jpeg


Here is the SpeedHut inverter driving the GlowHut panel:
1659141719073.jpeg


As you can see, the OEM inverter drives the panel to 295V P-P and the SpeedHut inverter drives the panel to 226V, a difference of almost 70V. This explains why clusters with the SpeedHut inverter are dimmer than ones with the OEM Ford inverters.

I then cut off the dimmer pot from the SpeedHut inverter to see how much impact getting infinite resistance on the dimmer pot would get us. That got the SpeedHut voltage up to 228V, just a 2v improvement.

Thought you might find this information interesting.
 

mwolson

Gray beard
Established Member
Joined
Mar 28, 2004
Messages
3,411
Location
San Jose, CA
Good question. I don't know. I have a theory that they suddenly fail completely but that is just a theory...
 

mwolson

Gray beard
Established Member
Joined
Mar 28, 2004
Messages
3,411
Location
San Jose, CA
I have developed schematics as far as I can without taking the components off of the PCB. The resistor numbers I chose may not match the numbers on the PCB since I can read them with the components in place. Here are the schematics as far as I have gotten: (EDIT: Updated to R2.0)

1661613147809.jpeg


Based on the signals I saw on the base and collector of Q1, it is an NPN bipolar transistor, not an N-channel MOSFET.

I am assuming that the transistor is the component that will most likely fail, but will have to see when I get a failed inverter. I have 4-5 candidate transistors that may work as a replacement part for the FST837 transistor. I will have to try them to see. I still have to get the components off of the PCB to figure out the transformer wiring, the part number of D1 and the value of C2.

I am still looking for a failed inverter to take apart and to try to fix. I also hear that 03 Lightning clusters have the same inverter in them...
 

Attachments

  • 1660365179331.png
    1660365179331.png
    215.8 KB · Views: 154
Last edited:

bull_dog190

Well-Known Member
Established Member
Joined
Jun 27, 2007
Messages
187
Location
U.P. Michigan
I have developed schematics as far as I can without taking the components off of the PCB. The resistor numbers I chose may not match the numbers on the PCB since I can read them with the components in place. Here are the schematics as far as I have gotten:

View attachment 1757948

Based on the signals I saw on the base and collector of Q1, it is an NPN bipolar transistor, not an N-channel MOSFET.

I am assuming that the transistor is the component that will most likely fail, but will have to see when I get a failed inverter. I have 4-5 candidate transistors that may work as a replacement part for the FST837 transistor. I will have to try them to see. I still have to get the components off of the PCB to figure out the transformer wiring, the part number of D1 and the value of C2.

I am still looking for a failed inverter to take apart and to try to fix. I also hear that 03 Lightning clusters have the same inverter in them...
bumping for a good cause.... anyone have a failed inverter to sent?
 

mwolson

Gray beard
Established Member
Joined
Mar 28, 2004
Messages
3,411
Location
San Jose, CA
Thanks to Cobra_Clark98 (from a different forum), who sent me his 2002 Lightning cluster with a definitely blown inverter, I have learned that the similar vintage Lightnings use the same inverter to drive their EL panels. He also shared some Lightning forum posts that showed me that what I had thought was a choke filter inductor is actually a soldered in 1/4A fuse.

The fuse is a Littelfuse R251.250 available at electronics distributors such as Digikey. This is what it looks like:


1661611224374-jpeg.1089940



According to the Lightning forum posts, many inverter failures are due to this fuse blowing. That can be easily fixed by simply replacing that fuse. The fuse is the first thing you should check if you suspect a blown inverter. You can do this check with an ohmmeter across the two leads. You should read very close to 0 ohms across the two leads. To make the test easy, I have annotated a photo of the bottom of the PCB:

1661611465156-jpeg.1089941



On the Lightning forum, someone suggested that the 1/4A fuse be replaced with a 1/3A fuse. I think this is a bad idea. The fuse is designed to protect the components on the PCB. The fuse is easy to source, but the other components on the PCB are not, so if they burn out instead of the fuse, then your inverter may not be repairable. I recommend that you stick with the 1/4A fuse.

This led me to compare the good inverter with the blown inverter. I was able to figure out which transformer pins were connected via windings and what the in-circuit resistances should be. The second test you can do is to check the resistances between the 3 pin pairs to see if the transformer is blown. Between Pins 2 and 3, you should see about 140 Ohms. Between Pins 1 and 4, you should see about 7 ohms. Between Pins 5 and 6, you should see about 12 ohms. Here is an annotated photo of the bottom of the PCB showing the pins you should test:

1661612151042-jpeg.1089942



With the blown Lightning inverter, the resistance across pins 1 and 4 was in the 10s of K Ohms, which means that the primary winding of the transformer was blown out. Interestingly, the fuse is still intact, which doesn't make sense to me. The 120 ohm current limiting resistor should have limited the current through that winding to about 100mA. And the fuse should have limited the current through that winding to 250mA, so I can't understand why it blew out.

Unfortunately, I have not been able to find a source for a replacement transformer. It is possible that Ford had a custom transformer made for these clusters. At this point I have no way of repairing this Lightning inverter unless I can find an off-the-shelf transformer.

The next question is, what are the other common failure mechanisms and how frequently do they occur? In order to answer that question, we will have to look at more failed inverters. So I am still on the hunt for failed inverters to analyze and possibly repair.
 

mwolson

Gray beard
Established Member
Joined
Mar 28, 2004
Messages
3,411
Location
San Jose, CA
Cobra_Clark98 told me that the reason the fuse wasn't blown on the Lightning inverter was because he had replaced it in an attempt to fix the inverter.
 

MalcolmV8

Well-Known Member
Established Member
Joined
Mar 27, 2003
Messages
7,353
Location
Tampa, FL
Hi Mark,

I don't have a blown inverter for you but I've enjoyed the thread. Great work and I look forward to what you figure out and what comes of this.

Malcolm
 

mwolson

Gray beard
Established Member
Joined
Mar 28, 2004
Messages
3,411
Location
San Jose, CA
If you don't want your car to be down while this is going on, you can take your cluster out, open it up, remove the inverter and put it all back together and back into the car. The cluster will work exactly the same without the inverter as it did with the inverter and/or EL panel already not working. Then the inverter can be sent to me to work on.

If you can find a reasonably priced working junkyard Cobra cluster, you can swap your cluster PCB into the donor cluster and be back to normal right away. Then you can send the dead inverter and/or EL panel to me to work on.
 

Users who are viewing this thread



Top