Why do I need an external amplifier keying relay?
Excerpt from FAQ: Icom IC-746 HF/VHF Transceiver (get PDF version)
Coordinator: Dana Hoggatt KB9SSS
Copyright (c) 2001, 2002 by Dana Hoggatt, all rights reserved.
This FAQ is also applicable to the IC-756, IC-756Pro, IC-756Pro II, IC-746Pro and all other Icom HF transceivers using the OMR-109F reed relay.
The IC-756Pro II User Manual, Page 18, re-states the SEND relay specifications as 16V 0.5A [non-inductive] - now that's more like it!
It is far too easy to overload and damage the internal switching relay, resulting in expensive and/or annoying repairs. The external isolation relays are cheap insurance for our rigs. Besides, even Icom technical support recommends them.
Given that the relay is prone to overload, the natural question that arises is; just how much can it handle? Things get both lengthy and muddy from this point on, but since you're still reading, we'll press on. Remember, you were warned.
This continues to be a process of discovery, so I'll present what we know in that form.
Page 79 of the instruction manual says that the SEND jack is rated for 16 Vdc at 2A. Some rigs have failed while well within those specs. Others seem to survive conditions far in excess of that rating. Obviously, the ratings in the manual don't tell the full story. (Editor's Note: Page 18 of the IC-756Pro II instruction manual, Rev. 2, states the SEND relay rating as 16V DC 0.5A.)
Homebrew explanations of (and solutions for) the problems abound. Some talk about bypass capacitors, while others talk about suppression diodes. Folks debate inductive vs. resistive loads and ratings. Surge currents and hot switching usually get brought up, as well as AC vs DC loading. For years, the topic keeps coming up, but with no consensus or dependable solution; except that using an isolation relay/circuit seems to prevent any failures.
So, many of us turned to technical support, but received different answers at different times. For example, in the qth.net Icom reflector archives for January 2000, we see that tech support told AB2BK that there was a misprint in the manual. The true rating was 12 Vdc at 20mA and Icom recommended an external isolation relay. While that would explain the failures, it makes some of the successes almost unbelievable. On the other hand, the archives for February 2002 show that tech support told N1JM that "The real scoop is UP TO 16V OR UP TO 2 amps.", again, recommending an isolation relay.
While the techs did not agree upon the ratings, they did agree upon using the external relay. Unfortunately, none of the answers from tech support have adequately explained the odd successes and failures.
All of this boils down to how much abuse relay RL1271 on the Main board can take. The parts list on page 6-12 shows the relay as an OMR-109F. The obvious solution is to look up the parts and see what they are rated for. As of January, 2002, nobody had revealed any spec sheets for the relay.
Searching for "OMR-109" on the Internet did turn up a curious tidbit. Icom has used the OMR-109 before, as the SEND relay (RL2) in the IC-761. It had problems too. There is a Service Bulletin (24287-001A) entitled "Larger relay for keying non-ICOM amplifiers" which recommends:
"Replace RL 2 (currently labeled OMR 109) on the PA Connector Board with a SY-12 relay (ICOM P/N 921-04696). This relay is larger but can be squeezed into place. Glue this relay to the board with RTV.
Check out: ic761relay.txt (Icom B2502)
More searching did turn up information for the SY-12-K relay, which lists a contact rating of 1A @ 24Vdc or 0.5 A @ 120Vac (resistive). We can infer that the OMR-109 specs are most likely less than these.
Fujitsu SY-12-K Relay Data Sheet
Unfortunately, we still did not have many answers.
Finally, in February 2002, we found a listing for the OMR-109F (thanks to the kind folks at Omron Electronics for helping out). It didn't say much, but did show a contact current rating of 1A. (Note: The http://relays.tycoelectronics.com/oeg.asp
http://relays.tycoelectronics.com/datasheets/OMR.pdf (Alternative Download) (OEG Data Sheet)
|Max. Switched Voltage:||AC: 120V.|
|Max. Switched Current:||1A (OMR-F)|
|Max. Switched Power:||OMR-F: 50VA, 50W.|
|OMR-H: 10VA, 10W.|
|Max. Switching Rate:||300 ops./min. (no load).|
|30 ops./min. (rated load).|
|Operate Time:||1.0 mS|
|Release Time:||0.5 mS|
|Expected Mech. Life:||100 million operations (no load).|
|Expected Elect. Life:||1,000,000 operations (rated load).|
First we see that the listed current/voltage ratings do not match anything we've seen or have been told before. Curious that. Why has Icom technical support given out at least TWO different sets of specs? One theory is that the relay specs we have are still not detailed enough.
It turns out that there are two ways to report maximum contact current. "Maximum Switched Current" is the maximum current at switching time; when the contacts are opening or closing. Once the relay is switched on, the contact current may rise to as high as the "Maximum Carrying Current", but must reduce again prior to switching off. Thus, if the maximum carrying current is 2A, then the answer from February 2002 could make sense.
Closer examination of the maximum switching rates reveals a possible problem. Suppose, for example, I transmit at only 1 word per minute. We average 5 characters per word, 3 dit-dahs per character, and 2 switches per dit-dah. So, CW at 1 wpm will generate 30 relay operations per minute. That's the limit of what the relay can handle at full load. 10 wpm will generate 300 op/min, which is the no-load limit for the relay. 30-40 wpm (what my local CW contesters do) pushes us to about 1000 op/min; which is well beyond the relay's rating. The max internal keyer speed of 67 wpm (as per ARRL review) would yield a brutal 2010 op/sec.
This could account for the varying results encountered in the field. We'd all assumed till now that the problem was simply one of current or voltage overload. Now we can see that those operating QSK CW are overstressing the part, no matter what the load. Pushing the part that far beyond its operational limits would certainly make it more failure prone, which could explain some of the failures that have happened within the supposed current limitations.
Consider this: CW at 20 wpm would generate about 600 op/min, which would reach the expected mechanical life of the part after less than 3000 hours of transmitting. But wait; if we apply that same rate to the expected electrical life, we get less than 30 hours of transmitting time. In contrast, AM/FM/SSB/RTTY operation would likely require 200,000 and 2000 hours of operating time, respectively, to hit the expected mechanical and electrical lifetimes. ( Note that reliance on these projected life specs can be very misleading. Read additional comments on this issue.)
The schematic on page 11-4 of the IC-746 Service Manual shows the SEND relay as an AHY103, not the OMR-109F listed in the Parts List on page 6-12.
(This is listed in the "Discrepancy" section of the FAQ)
Currently, we believe that the parts list is correct and that the schematic is wrong. However, this might help explain why Icom Tech Support has quoted different specs at different times.
We have now identified MANY possible failure conditions for the SEND relay.
These are the reasons for using an external SEND relay. Build Matt KK5DR's or Ted K9IMM's keying relay interface. Alternatively, you can buy one - from Ameritron or from Gary KD9SV. In addition, Harbach Electronics offers the SK-200 and SK-220 interfaces for the Heathkit SB-200 and SB-220. The AmpKeyer offers yet another choice. SultanTronics also offers several keying and receiver-protection interface products.
In some cases, it may be possible to tap the OMR-109F relay sharply with a small tool, to "unstick" fused reed contacts. This should be considered as a temporary measure until the relay can be replaced. (Contributed by Mel Martin VE2DC)
Read "Keying Amplifiers with an Icom Exciter". Click here for keying-line specs on various HF tube amplifiers.
Last updated: 11/27/2015