Continuing EMC issues
The May “Technically Speaking” column addressed electromagnetic compatibility problems with electronic control modules (ECM) and radio communication equipment within the confines of motor vehicles. A specific problem was discussed along with test procedures to determine its severity. This column continues that discussion and highlights what has been done since that time.
The problem involved International transport trucks used to transport forestry fire-suppression “crawler” tractors. To recap, the ECM that was used to control the automatic transmission (Allison) and manufactured specifically by Motorola is causing severe degradation to the VHF high-band radio (Motorola) installed in the truck. The truck is a 1996 model International 4900 transport truck with Allison automatic transmission-World Transmission model WT2 with shift lever control.
Another larger truck, also manufactured by International, exhibited severe interference to the radio. The interference locked the scanner on a specific channel even though that channel was programmed for a CTCSS tone squelch. The noise rendered the radio virtually useless at that particular frequency. Unfortunately, that frequency is critically important for that particular truck. The inability to use that frequency places severe limitations on the usefulness of the transport truck.
It turns out that this noise problem was not caused by the automatic transmission control module, but by a module controlling the running of the engine. The engine does not have to be running for the noise to be generated. Simply turning on the ignition switch starts the noise. The ECM, manufactured by Cummins, is mounted on the engine block. Removing the fuse to this controller stopped the noise. This was a step in diagnosing the source, not in solving the problem, because the motor cannot operate with the fuse removed.
Consultation with Motorola’s engineers raised the issue of just where the noise was entering the radio. Based on prior tests, it was my contention that the noise was radiated and picked up by the antenna and not conducted through the cabling. However, one of the engineers suggested placing ferrites on all cabling leading to the radio-coax line, power cable and speaker cable. This had minor results. The engineer’s suggestion was to do more of this with better ferrites.
The engineers left, the truck was returned to its normal location-and the noise persisted. Because the truck was located in a rural area where the local RF noise should be quite low, I decided to perform more on-site tests to further pin down the source of entry of the RF noise into the receiver. Could it be entering through the cabling, or was it entering through the antenna? This issue had to be resolved. The previous test procedure that was used to determine the amount of degradation caused by the electromagnetic interference (EMI) involved the use of a directional coupler connected as shown in Figure 1 above. In the absence of the EMI, the signal generator was set to produce 12dB sinad at the receiver output. Then, with the ignition switch on and the EMI present, the signal generator level was increased to again produce 12dB sinad. The difference between the two signal generator levels is equal to the amount of degradation caused by the EMI.
A new “non-invasive” test procedure was devised for convenience and integrity. There was no physical change to the receiving system cabling, so the test itself did not change the situation. Also, this was more convenient because the radio was mounted within the console above the windshield, and access to the radio was difficult. To get RF coupling from the signal generator, a magnetic antenna was placed near the radio’s antenna on top of the truck cab. With the magnetic antenna connected to the signal generator, the desired on-channel RF signal was coupled into the radio’s antenna and at the same time, isolated from it by distance, as shown in Figure 2 on page 57.
A speaker was connected to the input of the sinad meter and then placed close to the radio’s speaker. This acoustic coupling was sufficient to drive the sinad meter input. Double-sided adhesive foam helped to hold the coupling speaker to the radio speaker during the test procedure. Thus, the entire test setup was “non-invasive,” that is, the system cabling was not disturbed at all, as shown in Figure 3 on page 58.
First, with the radio’s antenna whip connected and the truck ignition switch turned off, the signal generator was set to produce 12dB sinad and the signal generator level was recorded in dBm. Next, the ignition switch was turned on and the signal generator level was increased to produce 12dB sinad. At one frequency, 151.400MHz, the signal generator level had to be increased by 11dB to overcome the EMI. At 159.420MHz, the signal generator level had to be increased by 7dB to overcome the EMI. The amount of degradation is frequency-dependent.
Next, with the spring and quarterwave whip removed from the radio’s antenna mount, the same test was repeated. This time there was no difference between having the ignition switch turned on and turned off. This indicated that the point of pickup for the EMI was not the cabling but the antenna whip itself. After all, the cabling was the same. The only difference was the removal of the antenna spring and whip from the radio’s antenna mount. Obviously, a much greater signal level was required from the signal generator to produce 12dB sinad with the radio’s antenna spring and whip removed.
Though we have been persistent in our efforts to get the manufacturers to fix these problems, to date they remain virtually unabated. However, some reduction was achieved on certain frequencies by changing the frequency of the microprocessor crystal. Sufficient relief is yet to be obtained. Again, I will keep you posted on further actions and developments in future columns.
* * * On the subject of vehicular EMC problems, I received a letter from H.M. “Jack” Tibbetts, senior manager of cellular systems for CellularOne in Burlington, MA. An excerpt from his letter follows.
Your article (May 1998) brought to mind similar experiences that I had. In the early years of vehicle ECUs, a customer of the service shop I was working for brought in his new Cadillac to have us install the latest GE Executive II 150MHz IMTS unit.
We installed and tested the unit, and he left happy. About a month later he came back for us to repair a small problem. We were talking during the course of the repair, and I asked him how he liked his new car. He lamented that he was having a lot of trouble with it and was tired of making repeated trips to the dealer. I inquired about the car problem, and he told me that the car would occasionally, for no apparent reason, run very roughly. The dealer was never able to find anything wrong but was on the verge of replacing the transmission under the vehicle’s warranty.
The dealer had involved the Cadillac zone manager for the Boston area in this problem, and they both had agreed to the transmission replacement. At this time I had not given any thought to a radio/ECU interference problem. We continued to discuss his car problem, and I held his IMTS unit at the shop to continue the repair. I let him leave without the IMTS unit and asked him to come back the first of the week.
When he returned to pick up his radio, I asked him how the car was running. He told me that it ran all weekend OK, and that he had done a lot of driving.
During the repair of his radio, I had to refer to the service manual, and I noticed that GE had included some verbiage relative to checking out the radio in the vehicle to make sure that there was no interference with vehicle electronics. I asked him to take me for a ride around the block while I made a test call on the phone. He agreed, and we set off with him driving and me making the phone call. As soon as the radio transmitter keyed, the car engine started to run rough, and the car kind of jerked along. I immediately terminated the call, and we returned to the shop.
He gave me the name and telephone number of the Cadillac zone manager. I put in a call to him, and we discussed the problem. After several minutes of discussion and questioning, we learned that the Cadillac engine ECU was housed in a plastic case for the first time that year. For some reason or another, GM had decided to try a plastic housing on the ECUs in their Cadillacs first. The Cadillac zone manager arranged to get a Chevrolet ECU (same unit only in a metal case) and have it installed in this customer’s car. PROBLEM SOLVED! The customer was happy, and Cadillac was happy. Customer: car and phone working perfectly. Cadillac: no expensive transmission replacement.
A similar incident happened with the Cadillac assigned to the president of Raytheon. He solved the problem by having his engineers wrap the ECU and its wiring in aluminum foil. His radio was a 110W VHF GE Executive II.
My worst nightmare is that the electronics in my vehicle are going to be interfered with by a radio transmitter in a vehicle next to me on the highway.
Thanks to Mr. Tibbetts for sharing this information with us. Problems such as this will get worse unless manufacturers attack this problem at design time. Once the RF noise gets out of the box, it is difficult to put it back. Often, simple solutions, such as the aluminum foil, just aren’t sufficient to suppress the noise below the interfering threshold. Manufacturers will only take heed when customers become more discriminate or when fleet managers, when letting bids for vehicles, include specifications that deal with EMC problems where it hurts-in the wallet!
Until next time-stay tuned!