Modification and verification of the FT-847 to optimize 70MHz performance
There is a lot of information on the internet about using the the FT-847 on 70MHz, unfortunately almost all of it is incorrect. This memo was originally produced in April 2000 for private circulation.
The modifications described here are valid for models with serial numbers "9E" (May 1999) or later. Prior to that date the firmware is at an earlier issue and has not been verified.
If you are unaware, YAESU date codes are the year of manufacture followed by the calender month as a letter code. As examples, FEB 1998 = 8B, MAY 1999 = 9E and JULY 2001 = 1G.
Note that serial numbers 9E and onward models have additional circuitry in the way of wire modifications added to modify the APC loop on TX. This is to reduce the output power to circa 15W between 54-76MHz. This was due to too many blown PA's.
Allowing the APC to be open can result in anything from 25W up to 40W output, albeit at >20A supply current.
1. Verify CPU
First you must examine the boot condition for the CPU. The correct product configuration codes for a 70MHz model are 0x2C or 0x6C(hex). This is set by 6 hardware jumpers adjacent to the CPU backup battery. The jumpers are numbered 1(LSB) to 6 (MSB). These settings are independent of the CPU reset function, therefore a full reset is not required after modification.
For code 0x2C or 0x6C the settings are:
1 2 3 4 5 6
0 0 1 1 0 X
0=closed (a short or zero ohm resistor link)
X=open or closed depending on wideband 37-76 TX or just 70.00-70.50 TX
Please check and confirm the configuration. The picture shows the battery and the adjacent jumper location:-
However, to be sure of the firmware code one must also verify that the serial boot EEPROM is correct. You will need to measure the voltage at several test points on the RF-UNIT. The filter select data is serially clocked into I.C. Q3018
When 70MHz is selected and the radio is on TX verify the following (refer to circuit diagrams for all references):
TP3012 = <0.4V
If the jumper selection at the CPU is incorrect then the wrong BPF will be selected. It may appear to work but very bad spurious outputs in the range 40-45MHz will be present when transmitting on 70MHz.
Models after 9E should be OK. Models prior to this require some additional switching added around Q3018 unless the firmware is reloaded.
See this page for further information on the UK import modifications.
2. Inductor L5006 modification
I have investigated the power output issue further after requests from local stations to modify some radios. Further examination reveals an apparent design flaw in the PA which I had previously overlooked but found by accident. This now permits a non-invasive modification to substantially increase the output power and efficiency, whilst the radio may be restored to its original state at any time.
Inductor L5006 (see circuit) which is part of an m-derived low-pass filter, has a rogue spurious resonance close to 70MHz. This can be easily verified with a simple GDO.
The modification moves the resonance down slightly by inserting a small ferrite core into L5006. This inductor is a 5-turn coil approx 7mm in diameter adjacent to the TX/RX C/O relay RL5001. (to the left of the relays in PA photo).
The core I used was taken from an old TOKO S102 series VHF/UHF coil and is held in place by a small dab of non-corrosive bath sealant, or acrylic from a hot-melt glue gun.
The power output of the unmodified radios was 20W. After modification, the power output had increased to 65W/75W respectively.
Some FT847s have a wired mod fitted around the "50" VSWR-FORWARD pot on the RX-IF board. This wire connection to the "50" pot can now be removed, as the ALC will limit at the nominal 100W 50MHz level). The limiting factor now is drive power as measured out of the exciter (AF-CTRL) board. The original current consumption at 20W output was nearly 20A. With the modified L5006, at 75W out it is now just 17A (There is still significant power lost in the main PA filters which have not been touched - see section 3 below for a modification). The power output at 50MHz remains at 100W, with a small drop in efficiency estimated as ~0.35dB loss.
After 15 minutes' continuous transmission at 100W on 50MHz, the modified L5006 core is just warm to the touch.
The second harmonic level at 50 and 70MHz is >55dB down on both radios tested, essentially the same as before the modification.
3. Low-pass filter modification
The output of the final stage is coupled into a LPF comprising C5057, C5058, C5061 and L5006. This filter has a loss of 0.9dB at 70.2MHz.
When operating on 70MHz the low-pass filter LPF is selected. LPF is a 7th-order Chebychev filter which has been tweaked to put two deep attenuation poles at approximately 90MHz and 110MHz (to cover the EU broadcast band). The component values as fitted on 70MHz models are:
C5123 33p fitted on at least 8G, 8J, 8K S/N models. Deleted on 9E onwards.
The filter characteristics are a compromise at 70MHz and are measured as:
50MHz: insertion loss 0.13dB; VSWR 1.12.
70MHz:(C5123 fitted) insertion loss 1.4dB; VSWR 1.2.
70MHz:(C5123 omitted) insertion loss 1.2dB; VSWR 1.4.
The total loss in the PA-UNIT output assembly at 70.2MHz is approximately 2.5dB (2.3dB) in the TX path and 1.4dB (1.2dB) in the RX path.
See also the supplementary information in the Equipment section of the Forum.
Between 54-76MHz, a broadband RF amplifier is selected. This has a noise-figure of approximately 11-12dB on 70MHz.
As a comparison, the RX gain is approximately 8dB lower on 70MHz than 50MHz. Note that increasing the RX I.F. gain via the menu will not improve the S/N ratio of the RX. It might sound healthier but do not be fooled.
Most later models have a small piggy-back PCB mounted on the RF-UNIT which contains an attenuator and LPF assembly for use below 1.82MHz.
This piggy-back board may be replaced with a switched 70MHz pre-amp, as detailed below.
Keith and Ian PA4ZP have now produced a miniature pre-amp PCB, which is similar in size to the existing daughter board mounted on the FT-847 RF PCB. The pre-amp board mounts in almost the same position but is fixed rigidly by soldering its mounting pins to two of the adjacent RF inductor sheilding cans. This is a minimal soldering operation to ensure the PCB remains adequately fixed.
The 3-way miniature power and control connector requires one wire to be soldered to ground (0V) and one wire to be soldered to a pad on the BPF diode selection matrix: this point is already occupied by the existing Yaesu 4m modification which remains intact (even if you have performed the power output mod).
The third wire is connected to +12V by means of a M3 eyelet tag to the existing +12V supply; this point is already occupied by the +12V supply to the existing daughter board.
The existing RF cable from the antenna changeover plugs directly into the pre-amp board, an output lead must be made up unless you wish to remove the existing cable from the redundant daughter board.
The pre-amp has a nominal gain of ~20dB, and its noise figure is <1dB. An output attenuator is provided: the values specified on the schematic are for ~5dB, however any value of attenuation may be used by changing the three resistors in the T-network.
The pre-amp contains a fixed BPF at the output: the -3dB bandwidth covers the entire OIRT band (66-74MHz). The filter BW may be narrowed if required, by changing the values of five components (details are provided).
In the bypass position, the HPF function of the original daughter board is duplicated using parts with superior Q (the original HPF has around 1.8dB loss at 50MHz, the new HPF is <0.4dB).
The PCBs are plated-through hole boards and are gold flashed to meet current RoHs manufacturing requirements. Assembly by hand will take an experienced SMD solderer around 35-45 minutes, but a first-time assembler should allow 2-3 hours.
The complete kit is available, including the CD, all mating connectors and cable to make up the co-ax jumper lead from the pre-amp output to the RF board.
Due to the increasing cost of components, the price including UK delivery is now £43.50 (as of April 2008).
Please note that this project is only available in kit form, so you'll need your own soldering iron too!
A second kit is under development which adds a small switched 70MHz BPF/Amp PCB at the TX output of the RF-UNIT at J3004. This addresses the very poor spurii from non-UK models in the TX range 54-76MHz when used on 70MHz. Please contact Keith if you are interested in this item.
For further information or ordering, please contact Keith G4FUF, whose address and email are correct on QRZ.com.