Racal Drivers unite

[WA1GFZ]

The typical Racal RX 6790, 1792, 6830 ETC all use an emitter follower between the output of the crystal filter diode switches and the input of the IF amp UA757.
I got my hands on a special second mixer module called very high dynamic range second mixer. The gain of this stage has been reduced and the power handling is up by at least 10 dB over the high dynamic range mixer module. I lose about 10 dB of sensitivity with this module. I tried gain stages in a couple different places and can't seem to get it back. I have run into this before but in stock radios there is plenty of gain ahead of the crystal filter module. A stock module does not have the close in performance.
I'm thinking of using a U310 common source with a tuned circuit in the drain and secondary winding to drive the UA757. the input Z is about 5 kohms. I need to get the mds into the crystal filter module to at least -115 dBM.
Stock filter module A10 MDS about -100 dBM. This modification would also help the stock second mixer by reducing the gain ahead of the second mixer which would increase close in dynamic range. So any Racal would could be helped. There is limited room to make this change so looking for suggestions.
This mixer module will process a 1 volt signal into the module without compression all the way to the 455 KHz filters. The first mixer of the 6830 just starts to compress at 10 db above that. (first mixer module /filter is passive with 10 db of loss) I'm shooting for -130 dBM to plus 20 dBM range of operation.
This will be enough dynamic range to put the phase noise of the synthesizer the limit of close in dynamic range. (then I will chase phase noise) I've measured 105 dB of dynamic range at 10 KHz spacing with this module in the radio. gfz

[w3jn]

It looks to me like it's almost a wash, Frank.  Lose 10 dB in MDS and gain 10 dB in dynamic range.  Couldn't you accomplish that with a simple 10 dB attenuator in the front end?

Perhaps that high performance mixer is faulty.... my  6790's first mixer shit the bed (one of the fets blew), lost about 10 dB in MDS. Iin the lack of a replacement I installed a SBL-1.  I now have a few replacements so I gotta delve into it some day.

73 John

[WA1GFZ]

John,
This mixer module came from a RA6840 DF radio. It used a different filter module with only 3 filters. The output has a low noise amplifier almost 20 dB.of gain to make up for the change in gain distribution. I need to make up for this loss of gain. This second mixer runs at plus 23 dBM LO about 10 dB higher than the common DF module. The common DF module gives me 100 dB of dynamic range at 20 KHz spacing in my hot rod unit. This new module I measure 105 dB (very hard to measure) But it is lacking in MDS by 10 dB. I will improve things if I can get the MDS back after the second mixer. I need to duplicate this low noise amp in the stock filter module to get the mds right again.
The cool thing is a stock radio could improve the close in dynamic range by doing the same fix because it will allow less gain ahead of the second mixer.
The LO of a 6790 is quite high that sbl 1 must be begging for mercey. A RAY6 or VAY1 would really be the one to use. SBL1 only needs plus 10 dBM I bet the 6790 lo is a half a watt.
I'm now considering duplicating that amplifier using a 2N5109 if I can get it to fit on the board in place of the piss ant 2N5087 emitter follower. It has 2 resistors in the collector lead so may be able to stuff a transformer in the holes. The cap off the emitter would go to the transformer. I was thinking of using a FEt because of the high input Z. The DF filter module LC transformer back to 50 ohms.

[w3jn]

How about a MMIC gain block?  Some have a NF of <1.5 dB, but I don't know if the IMD performance will be up to snuff.

Yeah I know the SBL-1 is singing for its supper, but it's all I had lying around.

73 John

[WA1GFZ]

John,
That was the first thing I tried. I'm losing too much mds through the filter section.
Looking at changing the switching diodes to pins as used in the DF filter module.
Presently 1N916. Trr is pretty fast so they may be lossy at 455 KHz.
I also cranked up the IF amp gain to a point of excessive noise. This module shaves the gain distribution for maximum dynamic range so the filter circuit becomes very critical. Maybe I need to install th emodule shield which could be effecting things.

[WA1GFZ]

John,
I was looking at the Harris RF590 schematics and sure enough they used a J310 FET as a source follower. I think I will give that a try since it is an easy modification. 1 cap and 1 resistor. Harris like Racal uses a UA733 to drive the crystal filters. The gain distribution very similar to the stock racal. Actually very similar circuits with the same U310 grounded gate amplifiers.  The Module I'm trying to get working replaces the UA733 with an ampifier similar to the RF amp used in the 6790 with a NPN driving a PNP. It will do about plus 20 dBM out...well above the 733. I don't want to amplify that any higher for fear of blowing out a filter. Harris also uses switching diodes so I may need to build a test fixture to see if it is worth going to pin diodes.
Harris uses a sbl 1 in the second mixer. I'm thinking the Racal could be easily modified to remove the MC1496 and switch to a DBM. The weak link in the Racal for close in.

[w3jn]

Man, I hate MC1496s with all those damn external parts.

I had some circuit boards that had a bunch of the smaller 4- or 5- pin version of the SBL-1 but I can't find 'em, those little mixers might be just the ticket.  They're about the size of an 8-pin DIP and maybe twice as tall, in a metal can.

Of course if you use a SBL-1 you're another 6 dB in the hole for gain.  This is between the roofing filter and the ultimate BW filters, correct?

[WA1GFZ]

John,
There is an LC filter ahead of the second mixer used in broad band applications where the filters are jumped out. I have a DF manual that shows a resistor in place of the filter to get very wide bandwidths. My present hotrod I pulled these filter parts out since I'm running 8 KHz. filters. This gave me the room to add AGC to the DF module.
I stuck a PAS-3 which has about 4 dB insertion loss.
The DF radios have a digital step attenuator on the fornt end and digital steps in the second IF. This sounds like a very cool set up but in reality the control signals inject clicks in the radio each time the attenuation changes. I built a controller once and it drove me crazy listening to the clicks.
Back to the subject. Anyway the 6830 has some reserve gain and the 4 dB available
be deleting the LC filter.  I have some Racal 1792/6790? engineering boards that converted the 1496 to a SBL1. The filter was also added. I think the 1792 only had one filter in the first IF.  The 1496 should be good for 90 dB of dynamic range if the gain distribution is right. This is an advantage of the Harris 590 using the sbl 1 I think. Racal makes up for it in the demodulator section though.
The 40 MHz in the 6830 arrives at the second mixer module at about 0 dBM so a little MMIC could be used to crank it up. The DF module uses a MWA130 to get it to about plus 14. The new module I'm playing with has an additional stage to get it to plus 24 dbM. Quite strapping for a second mixer. Both DF mixer modules have the same MDS. The newer one shifts more gain after the second mixer. I used a MMIC in the hot rod to get the mds to about -128 dBM.  Most stock 6830s end up with an MDS of about -132 dBM. Cranking it higher and you start to hear processor noise with the wider filters. Not bad for no RF stage. The R3030 I measure only 2 to 3 dB better with that strapping CP640 FET RF amp.
The Racal DF uses the RAY 6 mixer rated at plus 23 dBM. The RF590 uses the same type in the front end. SRA1H is not a bad one rated at plus 17 dBM LO. I bought a VAY1 form Mini Circuits promotion a awhile back and got a discount. It is a $75 mixer rated for a half watt. I really want to build an H mode when the first mixer becomes the problem stage. Who knows an H mode in the second mixer also sounds cool.
> Plus 40 dB IP3 and the fat lady sings.
It just hit me...Why does the MMIC work in the hot rod and not in the two other perfect working radios I tried the new module in??? The only difference is the shield is off the module and the stock DF I tried also had the shield cover removed. I need to check this. This DF module has EMI gaskets on the cover so it is very tight.
I could just be stupid here and missing a basic thing. The Second mixer must have a mds of -140 dBM  for the first mixer to get -130 dBM. that is quiet. R390A quiet!

[VE7 Kilohertz]

Hi Guys,

How timely this discussion is as I have my 6790 apart on the bench for an alignment and a few simple mods to the AGC and audio gain.

What are your thoughts re: making the detector true synchronous? i.e. phase locking the 455 out of the 1357 into the 1496. I have some cool little ADI sample boards with the AD607 on it. They would make great stand alone synch detectors. I also was thinking about just putting in a Philips 567 PLL or some other simple 8 pin DIP to phase lock the limited IF into the 1496.

Your thoughts?

Cheers
Paul

[VE7 Kilohertz]

One more thing.

I have some spare boards for the 6790 series and one of them is a fully stuffed with Collins mechanical filters Main IF board. I was thinking of trying some of these filters in my 6790 that has crystal filters but I see that the matching networks are somewhat different. Have you tried mixing the mechanical filters in with the crystal types? Which type of filter do you prefer? Xstal or mech and why? Do you have any crystal filters to sell ot trade in the 6-8 Khz range?

Paul

[WA1GFZ]

Paul,
No problem mixing crystal and mechanical filters. I do it all the time. I actually have a special 6830 board with an AUX 15.5 khz wide if strip that uses 2 crystal filters and 4 mechanical filters.
The mechanical filters require caps on the output and input of each filter. The mechanical filters have a bit less loss and a bit more passband ripple. Both work fine.
I took an FT1000 board and made a locking I/Q demodulator for a DSP front end.
I have not tried the 607 but the stock Racal carrier recovery works better. Look at the lo injection into the 1496 from the 1357. This is a cool design that doesn't have to acquire a phase lock. fc

[W3SLK]

Cripes, I wouldn't mind just having a RA-17

[WA1GFZ]

Well I have some loss numbers. The Racal filter board same circuit as the 6790 loses 23 db MDS through the passive power splitter/matching, filter selected to the emitter follower output. The MDS of the IF AMP UA757 (50 ohm source) is -110 dBM through the radio. This makes the mds into the filter section about -87 dBM. I usually shoot for a -130 dBM MDS at the radio input. The first mixer module of the 6830 is passive loss at 10 dBM. I think the 6790 may have some gain just don't remember.
The case of the 6830 about 53 dB of gain is required in the second mixer module.
(the module tune up procedure shoots for a bit more)
The Harris 590 design is very similar using mostly the same parts.
This explains why so much gain is required in the second mixer module.

The special module I modified to work is lacking 20 to 25 dB of gain. The good thing is I can get the MDS to -130 dBM by adding gain ahead of the filters after the module. This will not degrade the second mixer dynamic range. It should increase the dynamic range by adding gain after the module as long as the 455 KHz filters don't  generate distortion. (since mds will decrease)

The output of the new module will generate about 1/10 watts about as hard as I want to drive the filter module. now to come up with 25 dB of clean gain that won't blow out the filters or make the filter section more efficient like the DF module with only 3 filters. 

Its going to be a long winter..... can I hit 100 dB dynamic range at 5 KHz spacing?

[w3jn]

THat's a good trick, Frank - you're sure not gonna be able to measure that due to filter bleedthru and/or skirt attenuation 

You might try a 7360 or 6AR8 sheet beam toob mixer, the dynamic range of these things is pretty incredible.

[WA1GFZ]

John,
I hope to do some dynamic range checks tonight. The trade off will be the filter dynamic range VS phase noise. I'm thinking of a voltage amplifier ahead of the filter module with plus and minus 15 volts across tha amplifier. A 50 ohm resistor to ground at the input. 2N3375 stud mounted bipolar running at 100 ma should do it.1 Kohm collector resistor should be a good source for the filter input. I guess I should put some sort of a limiter ahead of the filters to protect them. It seems the dynamic range of the mixers is no longer an issue. I hope the ability to measure dynamic range will become an issue soon. There is always the H mode mixer to explore if I get lucky IP3 plus 44 dBM.  fc

[VE7 Kilohertz]

Wow! You fellows are talking about 3dB above my level of receiver knowledge but I slowly learning.

I feel lucky to have completed my 6790 alignment.     I went thru the A4 IF alignment about 6 times before I got it set where I want it. The 2 different techniques outlined in the 2 docs you sent me Frank yield totally different results. I also modified the AGC per Dallas Lankford's Nov. '04 article but I think I need to modify the release of the MED AGC as it is too slow for fast break in QSO and I don't like the SHORT characteristcs.

Now I am moving onto my RA1792 which I discovered has the 100KHz IF board and the RF preamp. Cool! Also is loaded with STC xstal filters which seem to work very well. I may try an alignment on this one as well as it seems a little insensative, but shouldn't be, especially with the preamp.

Has anyone figured out how to retrofit some LED backlighting to the older Racals? I would love to get backlight into the RA1792 and the Collins 2050. Nice RXs but you can't read the display in a dim shack with a dim operator.

Cheers

Paul
VE7KHz

[WA1GFZ]

Paul,
 www.gewingerd@netzero.net he had some led boards. He is Mr Racal parts.
Let me know if you find a good procedure for the AGC. I still think it needs work.
The fast agc is only good when there is a lot of static. The quick release helps a lot.
You get a noise blanking effect.
Glad you fixed your 6790. Dallas is a sharp guy I like his fix for the audio level. the 6830 has a level pot that I found likes to be set lower for better audio quality to keep the final audio amp out of saturation that caused a weird phase distortion pumping action.
I would home brew the second mixer board if I owned a 6790 or 1792. There is a lot of performance hiding in that radio also. The synthesizer can be cleaned up as you must have read fron Dallas. I like the 6793 the best. Perfect match for the MA2232
spectrum monitor. (also good looking with the 1792) fc

[WA1GFZ]

Total JS set up gave me 110 dB dynamic range at 100 KHz spacing and 100 dB at 20 KHz. I'm using 20 KHz filters in the first if so close in it was only 80 dB. MDs=-130 dBM with 2 stages of gain between special second mixer and filter board.
I'm using a filter board with an aux 455 IF strip with an output to the rear panel.
The aux IF amp is a mc1590. Here is the kicker it has an mds of -130 without 2 stages of gain ahead of it. the aux stage has the same emitter follower after the filter as the main channel.
this boils down to two components 1. the switching diodes 1N914. maybe pins would have less loss. 2 the IF amplifier noise. Nothing was gained with an amplifer added ahead of the ua757 so the next thing to try is replacing the high speed diodes with pins. A stock radio 6830, 6790, 1792 Harris 590 all have the same high speed diode switches. A pin with a slow Trr should have less loss than a fast switching diode.

[VE7 Kilohertz]

Cool work Frank. Keep us posted on the diode change over. I have a variety of PIN diodes here that I used in my TS-830 mods a few years back. Still have some. I also used them in the detector in my TFT AM mod monitor.

Hey, do you or anyone else here have a copy of Dallas Lankford's noise Blanker cct for the 6790? I can't find it on the web. Would like to try it.

Cheers

Paul

[WA1GFZ]

Paul,
I would like to see Dallas's noise blanker so send me a copy if you find it. I have not heard from him in a while. I guess the wX down south may have got to him.
Last night I down loaded a pile of pin diode data. I'm thinking of using the MaCom pins used in DF boards. I also looked at a 6830 DF manual today and the 3 filter board has only 3 DB of loss not 23 dB. So there is my missing 20 dB of MDS. I want to reduce the loss in the filter module before adding a lot of gain ahead of it that will limit close in dynamic range. I found some cool pin diode limiter circuits for when I add gain. The problem with hot rod receivers is power handling. I will have to protect the filters if I add a gain stage capable of 1 watt. Today was Transmitters and antennas

[VE7 Kilohertz]

Frank,
Check out HP/Agilent (now someone else) 1N series of PIN diodes. 1N5819 I think is the series. May be too pedestrian for what you are doing but I think that is what I have used.

I tried to find Dallas's email today and found one from LA Tech but I think he has retired?? Do you have a current email for him?

Cheers

Paul

[WA1GFZ]

Hi Paul,
Yup LSU for Dallas. I have not heard a peep frpm him. I'll ping him in the morning if I think of it. His email is at work. Well I just found a problem with my limiter module ahead of the first mixer. Seems I lose some dynamic range through it. The limiter kicks in at 5 volts and I am no where near it. The special second mixer is perfectly matched to the first mixer. The spurs come out of the noise at -20 DBM input with a noise floor of -130 dBM, 100 khz spacing. No strapper is that strong at my house.

Yes, I have all the pin stuff you talk about and more. I found some cool limiter and attenuator circuits. I read the pin diode needs about 10 ma to switch to a low resistance. This may be a problem in the 6830 filter module.

[WA1GFZ]

The difference between high speed diodes and pin diodes as switches is the current required to pass RF. A high speed diode requires the forward DC current greater than the RF current or it will switch off during part of the cycle causing distortion.
A pin diode will pass more RF current than the DC bias current due to the long Trr. 
Racal uses 1.5 ma of current so I need to look at what happens when this new module is wide open. It will produce over plus 20 dBM into 50 ohms. The Filter module has an input Z around 2 Kohms. (6790 a bit lower since it has more filters) The voltage swing is a lot higher at light load. The output stage of this module is very similar to the 6790 RF amplifer but set to higher gain. This makes me suspect these diodes are being switched even though they are after the filters. A filter will take plus 10 dBM with 10 dB loss says the diodes must handle 0 dBM. .225/50=I or around 4.5 ma RMS. The racal DF boards used higher switching currents but they are operating at 50 ohms.
So I would think these diodes distort badly when there is a strong signal using this module. I don't know about the stock UA733 in the 6830, 6790, 1792. I need to look at it.
The pin diodes seem to have about 50 to 100 ohms at 1.5 ma. This may be ok in the high Z circuit. They should work better. I'm looking for the data book that is hiding in the shack that covers the diodes used in the Racal DF receivers. They are used in the front end attenuator and 455 IF strip. They should work great. I just need to verify the series resistance at 1.5 ma. I will be dropping in one as a limiter ahead of the filters just in case. Also might try them in the front end limiter.
I ran into a number of articles on pin diodes as switches and attenuators. I found all the old HP stuff. Somewhere in the house I still have those books. racal used microwave associates pins. They are bigger than the HP parts. It looks like they will fit in the Racal boards with high speed diodes though.

[WA1GFZ]

Dynamic range of RF into the filter module 10 KHz spacing is 80 dB (Racal spec for RX at 10 KHz is 80 dB. I would think it could be a lot higher. Filters should be about 20 dB higher. I will start wit 1 filter installed and the switching diodes removed and replaced with shorts and opens. I think the 6790 has similar performance.

Third order intercept of stock mixer looks to be plus 41 dBM. pretty kick butt.
now to make the whole radio as good.

The best way to modify a radio or design one is start with the speaker and work forward.

[VE7 Kilohertz]

Hi Frank,

Thanks for keeping us updated on your progress. When we are done here, I am going to print this and keep it in my Racal reference file.

I dug thru some of my boxes last night and found some more of the PIN diodes I was playing with a few years ago. These ones are HP 5082-3080 and 5082-3083 and a couple others in that series. This was back in the days when I was in the electronics distribution game and samples came by the bushel.    I know I also have some mini ccts stuff and will try to locate tonight. Not sure what I have but they are 3 port little metal boxes. Maybe I'll find an early Christmas present. About 10 years ago I found a local surplus place selling 1000's of pounds of old cel phone gear (central office type stuff used by the telco) and it was loaded with gold plated SMA, SMB etc jumpers with 1' of silver plated coax, attenuators, N connector jumpers etc that just had to come home with me. A  ham's goldmine. All of this has been in storage until a month ago and now I am in heaven. When I did the Racal alignment I had all the push on plugs to go straight onto the Racal boards. Made testing a dream.

Cheers

Paul