Kikusui KSG4100 or Marconi TF2015 sig gen.

[Roger-M0TDM]

Title says it all. Which in your opinion would you have?

Regards
Roger.

[karesz*]

Hi Roger,
If for you is max.110MHz OK, i would take the Kik, but the other is a nice vintage thing too, then all things are relativ ...
K.

[dave cox]

There is a service manual available for the marconi and at least some of the parts are still available. The min freq (10MHz ?) of the tf2015 could be limiting unless you have another generator.

Good luck if you ever have to find a service manual for the Kikusui, otherwise its probably spot on for setting up domestic AM/FM and SW radios

[G0HZU_JMR]

I've got a TF2015 here and I've had it since the 1980s. It's proven to be a tough little generator and the only fault was when the tuning indicator cord snapped so I had to rewind it using lacing cord.

The bad points are that it has very poor phase noise and obviously it isn't synthesised so it is prone to drift unless it has had 45mins to stabilise.
Also you need to hook a counter to it as the dial accuracy is very vague and of limited use. However, you do need to fit an attenuator to the counter port or the leakage from this huge signal can leak into your radio under test if you have the generator output set near the lowest levels it can go to.

It should prove easy to find parts for as it uses quite a few classic components in the circuits. It also offers a sweep function. It's small, light and portable and has nice touchy feely controls. Great for servicing CBs and 6m/4m/2m gear. It can cover 70cms too but can get a bit drifty up there...

However, it only goes down to 10MHz so I made an external downconverter for mine to get it down below 100kHz. I rarely use the 2015 these days as I've since bought several generators that are fully synthesised.

The Kik sig gen looks nice but it does look to be designed for setting up domestic radio receivers so it might not be that great for CB/ham radio servicing etc.

i.e. the FM modulator may well prove to be optimised for high deviation and may prove to have dodgy accuracy down at narrowband settings.

Also you need to find out if the Kik has any reverse RF power protection. The little TF2015 has a diode limiter to protect it up to several Watts reverse power. That's partly why it's output only goes up to 0dBm because any higher and it would start to hit its own protection limiter LOL.

If the Kik has no protection then it would be very easy to fry the attenuators in the Kik if the transceiver under test went into Tx mode. I'd expect that replacement attenuators are as easy to find as replacement roof panels for Noah's Ark...

Note that you could easily make an external protection limiter for the Kik using some back to back diodes eg 1N4148 and this would provide pretty good protection up to 5Watts and would usually survive 10 or 12W on a short transmission. But the limiter would only allow generator outputs up to just over 0dBm before the limiter clipped the generator output itself!

Also it looks like it only does 1kHz tuning steps above 35MHz. I suppose this should be OK for most applications though

[Roger-M0TDM]

A very comprehensive explanation Jim thankyou...
I've passed on the kik,so it looks like the Marconi for me,and as you say it will be used for servicing cb's and my ham radio gear I have an old counter a phillips pm 6622 which suffices for my needs so I will check the correct carrier freq of course with thator hook the gen into my Tek scope and use the external socket to feed the counter decisions decisions and I welcome suggestions mate.
Regards Roger-M0TDM.

[Dickie]

Look for a TF2016. Similar to a TF2015 but goes down to 10KHz.

[karesz*]

Hi Roger,
>Check please the "TF2171 digital synchronizer (fro TF2015)", its some time to buy on epay etc...
Other option can be an TF-2018A/19A AM-FM Synthesized Generators up to 520/1040 MHz too?
K.

[G0HZU_JMR]

The Marconi 2019 is a very nice signal generator to use. I've got a couple of them here. They aren't very reliable over the long term though

I've used plenty of them over the years at work and they all proved unreliable. Mine are OK now but it took several years of fixing intermittent faults and the odd component failure to get them reliable. The attenuators can get sticky as well if not used regularly. So I'm always cautious about recommending them. However, if you get a good one and can afford the space then they are a very nice signal generator.

As Dickie suggests, the TF2016 does look to be another option. I think it covers 10kHz up to 120MHz.

To get operation below 10MHz I built a simple downconverter for the TF2015 using a mixer, a 30MHz xtal oscillator for the LO and an amp and filter.

This gave it unity conversion gain and meant that 30.010 to 39.999MHz from the generator mixed down to give 10kHz to 9.999MHz at the IF port of the mixer. This worked very well indeed and many modern generators use this approach to get operation down to very low frequencies.

Note that there are a couple of versions of the TF2015 because one version has the attenuator scaled in uV through mV and the other has the attenuator scaled in dBm. It may be the same for the TF2016?

[Roger-M0TDM]

Hi many thanks for the replies...
There are some 2019a on ebay but located in Israel so not cost effective.
Forgot to add I also have an eldely Marconi TF995A/2 in very good condition that has a good freq range,I also have the Pye series of sig gens for vhf and uhf ,a marconi tf2005 an old levell osc a feedback function with sweep and the old bc221 and the uhf version as well...
Have so much test gear its hard to remember what I dont have.
Regards Roger.

[karesz*]

Hi Roger,
your welcome, but what about an SMD?-their are very professional solutions-OK, not miniature...
K.

[Station X]

One post moved to a new thread here:-

https://www.vintage-radio.net/forum/...d.php?t=107789

[Tony Jaques]

Hi G0HZU,
I would be very interested in any further information on the down converter that you made. My TF2016 oscillator has become increasingly intermittent, and after several goes at it I have failed to cure it. But I also have a TF2015 so your down converter looks like a good cure for the '2016!
Thank you,
Tony (G3PTD).

[G0HZU_JMR]

Hi Tony
I'll dig it out and have a look inside and produce a basic block diagram and show a few photos. Sadly, I didn't produce a schematic at the time. It was built in a hurry using salvaged parts so it is a bit ugly inside.

Note that this downconversion method does introduce some undesirable side effects. Apart from the introducing low level spurious terms the downconverter will introduce a raised noise floor because it will have a fairly high noise figure.

I'm guessing (from memory) it will be close to 15dB and this compromises the integrity of the generator if you want to produce signals down below 1uV. This is because the downconverter will produce noise about 15dB above thermal noise depsite it having unity gain from input to output.
So it will act a bit like a sig gen together with a low level noise source. The added noise only really matters on tiny signals though.

I don't know if this is enough to put you off but I would recommend fixing the 2016. I made this converter out of desperation to get operation below 10MHz because I had no other (financially viable) option at the time.

It would be possible to reduce the overall noise figure of such a converter by changing the block diagram a bit. i.e. my design wasn't ideal. The other alternative would be to add a step attenuator on the output if you want to produce tiny signals.
i.e. fit a decent 20dB attenuator on the output and set the generator to 10uV if you want 1uV from the downconverter. This would give a better S/N ratio for the wanted 1uV signal at the downconverter output.

I also recall that I built it in an evening using dead bug style and put it inside an old salvaged die cast box that already had BNC connectors on it.

It was a very useful thing to have at the time because it allowed me to generate modulated signals down at 455kHz and also to do basic tests on receivers across the whole of the SW band

[G0HZU_JMR]

I found the old downconverter and here's a block diagram and a simulation of the linearity using a two tone test and also an image of the downconverter itself.

It's designed to cope with up to 0dBm PEP at the input and it has unity conversion gain. The output 3rd order intercept is about +16dBm both in the simulation and also for the actual box itself so even at 0dBm PEP the IMD distortion terms are better then 40dB down.

I've placed a few labels inside the box partly to show where everything is and also partly to hide just how ugly it is in terms of construction I really did fit this into that box in one evening using an old scrap piece of PCB and I used a scalpel to isolate the only track on the PCB for the power rail. The rest is built dead bug style. Fast and ugly, but it works OK. I must have built this about 25 years ago.

Back then I had to make the 30MHz LO from a discretely designed oscillator followed by a 2N2369A based tripler and a BPF and then an MSA0404 LO amplifier to get +7dBm. A modern design could use a cheap 30MHz oscillator and a buffer amp to get +7dBm at 30MHz. This would work better too.

The amplifiers in the signal path are MSA0304 and there are loads of modern equivalents. eg MSA0386 is a similar device in SMD. Or you could make the equivalent using a couple of NPN transistors as these devices only need to work up to about 10MHz.

The 7th order elliptic LPF uses Micrometals T37-6 (yellow) cores probably because that's all I had back then.

It really is a simple design and I can produce you a schematic if you want one.
With modern coupling caps it should be possible to get this to work right down to audio frequencies at the output with less than 1dB ripple/flatness across LF to 10MHz.

[Radio Wrangler]

One Marconi sig gen of that era which is not to be missed if one ever crops up is the TF2017

This is uses a cavity oscillator rather like the HP 8640, and has very low phase noise. It has several more creature comforts than the 8640 as well.

David

[Tony Jaques]

Hello again Jeremy,
I apologise for not replying sooner. Some medical problems have taken part of the time, but the main reason is that you prompting me to have another go at the '2016 stirred my conscience up. I really think that I have fixed it this time - but then I have thought that before!
Briefly though, because this is rather off the topic of signal generator comparison, I do appreciate the effort that you have put in to provide the info. Thank you very much. Even if the generator doesn't fail again your effort will not be wasted. Years ago I made G4PMK's "Simple Spectrum Analyser" (Radcom) and I fancy making a tracking generator for it. Your converter design could be a good start on the path.
Thank you Jeremy.

Tony...

[G0HZU_JMR]

Glad you got the 2016 working. I think that is the best outcome

I'm tempted to mod my converter to fit a 30MHz OCXO to it as I can also use it with my newer sig gens. These only go down to 250kHz and the converter can be used to get them to work right down to AF frequencies. I've got a pile of fast warmup 30MHz OCXOs (somewhere) here with very good specs and this would improve the accuracy of the 30MHz LO and would also improve the spectral purity.

Whilst investigating the OCXO mod I noted that I also got the block diagram slightly wrong as I actually used a 15MHz crystal and a doubler to get 30MHz rather than 10MHz and a tripler. It was a long time ago and I must have forgotten.