Oldish signal generators that can do 10-100Mhz square, sine, pulse signals

[maiakaat]

Hi

My second topic, I have been put off many signal generators that generate AM/FM signals, that go for lowish prices, so I was wondering if anybody knew of relatively simple reasonable quality signal generators that can do these frequencies for simple testing for under $100 (older second hand instruments) PWM would probably be useful. I thought a list of good low cost test instruments might be useful, as many of the new/second hand signal generators are effectively useless at less than 10Mhz to most people doing highish speed digital electronics I would have thought.

A pulse generator with a very fast rise-time would also be useful for a similar or slightly lower price

After registering and looking around the forum I cooked my dinner three times over :O(

[maiakaat]

"But then i hope I misread your message. you ask for 100MHz, digtal signals, DC and 40-50 pound in one sentence, to speak with Airosmith: dream on   And why you want to measure the current of a digital signal ? To measure the powersupply you are only talking DC. So why the 100MHz question.

Contact Keithley or Agilent, if you pay enough the maybe develope one for you. A TRMS meter with very high crest factor and a 100 MHz bandwidth  will cost a fortune. "

I should clarify this...

I am looking for any older second hand signal generators that can cover from 10Mhz up to 100Mhz approximately, or the upper end of this range.
They only need to do a sine wave and a square wave, although a pulse would be nice

Even the Arduino can do 16Mhz or so (I think) so what use is a 2Mhz Signal generator these days, such as many of the low end and Chinese ones around, or the 20 year old equivalents that go for under £50.

I have seen RF generators go for less than £100 (close to £50) although I don't know whether they can send an unmodulated signal, as I am just looking for a simple signal, not an AM or FM radio signal, some of them go over 1Ghz, I don't know how they differ from a 2Mhz signal generator for them to sell at such a low price.

TRMS is really irrelevant, its for testing quality of ADC circuits, and digital signal quality, propagation etc, I don't intend on working with power beyond the essentials.

If there is no such thing (the Ebay ones are not capable of a simple signal) then I will think about the £100 to £300 range, but Ebay is filled with crap Chinese products or very expensive second user equipment, and in-between is a lot of older and not-so old RF equipment, so advice on the usability of these for simple signal generation is probably useful for me and other newbies.)

I don't believe that the only options are custom builds from Agilent, or £1000 units, but then my posts were not so clear.

As a side note, I've seen a colour 20Ghz scope with 18Ghz trigger go for £770 combined, and a 1-3Ghz colour CRT scope go for £350.
If you are patient and diligent Ebay can offer some unbelievable bargains.

I know a lot of people probe forums (and other companies and forums) for business, but in software/server hardware they are often 3-400% more than a typical Ebay price in some cases, hence why the sites often do not list the actual price.)

[free_electron]

If you are going to be testing adc's you will need a spectrally pure generator with very low thd .... Ain't gonna find that below 2k$ even old and used...

As for the pulse you need a machine with fast rise time. Pulse delay is not important.
I would look toward a 8110 style machine or 8112.

DO NOT BUY A BROKEN 8112 ! 99% of the broken ones on ebay have the shaper hybrid shot and that is unobtainium....

[nctnico]

I have a working R&S SMDA with manual and schematics (in German) which could use a new home...
http://www.radiomuseum.org/r/rohde_am_fm_messsender_smda.html

[maiakaat]

@free_electron

Thanks, that's really good advice.

We really wanted to save that sort of money for an upgrade to the Rigol DS1102E. I have seen some options for around £600-700.

I was thinking of getting the equipment to test basic functionality in the near future, and then getting better quality instruments in a year or so.

Does this mean the RF signal generators can often be used without any modulation for a pure input signal?

I was thinking of using a home made pulse generator and signal generator in the mean-time just for very basic learning and testing (over the next month or so)

[maiakaat]

I noticed this add FS: HAMEG Modular System Series 8000 Bench Multimeter & Function Gen PRICE DROP!, I am in the UK though, so postage is high, and if there is a problem its practically impossible to sort out, not to mention the risks involved in buying from a forum

[maiakaat]

@nctnico

We are in the UK, so postage is likely to be very expensive

[nctnico]

Postage from NL to UK is about € 41 maybe cheaper if the manual and the box are lighter than expected. I've used the SMDA primarily as a sine wave generator. The output is supposed to have low noise and distortion because it is intended to be a transmitter for a wide range of frequencies.

[saturation]

For simple testing, look at the thread on the Hantek DDS 3x25.

[T4P]

Placing mechatrommer's GolTek unlocks 100MHz sine wave ...

[Mechatrommer]

you are dealing with devils if talking about cheap and > 20MHz stuffs. getting good signal and measurement is a pure luck or delusional. my 2cnts. YMMV.

[vk6zgo]

There are many Signal Generators which can do better than 100MHz with sinewaves,in fact,500MHz & above is common.
These do not do this frequency span all in one range,so each range can be optimised.

Square waves at 100MHz imply  a frequency response in all the internal generation equipment & amplifiers, flat from 100MHz to around 500MHz to produce a reasonable square wave at that frequency.

Wide range Signal Generators as used in RF work,do not usually have square waves as an option,as is common in many "Function Generators",as they are used for different purposes.

There are Generators which are designed to produce square waves with fundamental frequencies in the 100MHz region,but they are either special orders & very expensive,or small volume lines,& again,very expensive.

[nctnico]

Not really. Just get an ADF4351 based frequency generator board. It will output square waves with harmonics into the GHz regions.

[free_electron]

I would look for a used HP 865x type generator like a 8656b or 8657.
These have true sine oscillators and are controlled by a fractional divider pll system.
These generators easily do 900MHz and are very clean at the output. Can be had cheap.

Better would be 8647 but those are pricey....

Using awg's is not a good idea... The dac in those things is 14 bit at most.... Good luck doing distortion measurements on a 16 bit adc with a 14 bit dac ...

[vk6zgo]

Not really. Just get an ADF4351 based frequency generator board. It will output square waves with harmonics into the GHz regions.

a

It's a board!
A real Signal Generator is a lot more than that!

[maiakaat]

I've got a couple of AD chips, a pre-assembled AD9850 for £4.00, and a low spec AD9832 TSSOP from Farnell, but eventually - this year, along with a slightly better scope we are going to need a good signal generator, at least I know where a significant chunk of any money should go.

The project is first a bottom of the range singal gen/oscilloscope/logic analyzer for the Raspberry Pi (but useful not cheap crap, properly researched design) followed by a better design about 12 months later. The design is open hardware and open software. So it gives you an idea of how the instruments will be used.

Mostly I am in the theory at the moment, so there is no absolute need for an instrument, and I am good at the Ebay waiting game, my background is in software engineering so forgive the vague terminology, I am learning though. I have access to some research libs, but lack of access to IEEE is a downer, along with Wiley/Elsevier, who have better material than ACM (of which I am a member) and the open research on the web. My strength is in the software, but I'd also like to look at ways to innovate, rather than just patch up the weaknesses of such a hardware design.

The sign-gen could be a lower range variant, say 1-25MHz or something for now, for the low hundreds of pounds if available. We wanted to save for a Agilent 3000 series MSO, but it looks like the signal generator in the high hundreds is more important when we do get a new instrument. So going up to around £600 is not impossible, although it will not be for a few months.

Would an older analog gen of a similar specification be better than a newer signal gen in the £200-£500 range for this sort of task?

On the pulser, I'm not confident enough yet to make a good pulser, although I am aware of a few circuits out there, I also have bought a couple of comparators, the LT1721, and a Maxim device that I manage to get eventually through Farnell. Above 50MHz things get much tougher, so I want to learn more before delving into the design of higher frequency stuff, I appreciate it's really challenging some of the engineering at high frequencies. The pulser is purely to measure the rise time of a circuit (or components for suitability) I am aware our current scope limits the performance of a good pulser significantly.

The suggestion of the RF generator needing a new home has really interested me, although I don't know if it could be used for all tasks, mainly for sending in an analog sine to help with learning and basic design testing. I have no idea how accurate these older instruments are, and how well they hold their quality over time.

I'll post some stuff I am keeping an eye on for others who might be looking, despite if I go for any they will push the price up. Some feedback would be interesting too, this is not a endorsement of the seller of condition of the instrument. I've noted in the US there is a lot of instruments sold without warranty, or even a confirmation it works, I think in the UK laws are a bit tougher to get around, sellers are less keen to auction too, with fixed prices.

Advantest TR6846 bench multimeter - Auction - £20 so far
http://www.eBay.co.uk/itm/330775739799?ssPageName=STRK:MESINDXX:IT&_trksid=p3984.m1436.l2649#ht_500wt_1366


(unrelated HP bench PSU)
http://www.ebay.co.uk/itm/290759974129?ssPageName=STRK:MESINDXX:IT&_trksid=p3984.m1436.l2649#ht_500wt_1413

HP 8112 pulse generator - Auction £60 approx (a comment earlier warned of faulty units being sold)
http://www.ebay.co.uk/itm/320962738226?ssPageName=STRK:MESINDXX:IT&_trksid=p3984.m1436.l2649#ht_859wt_1401

R&S signal generator - £100 so far at auction
http://www.ebay.co.uk/itm/221104242484?ssPageName=STRK:MESINDXX:IT&_trksid=p3984.m1436.l2649#ht_890wt_1401

Racal Dana - £265 at auction - synthesizer
http://www.ebay.co.uk/itm/110934785901?ssPageName=STRK:MESINDXX:IT&_trksid=p3984.m1436.l2649#ht_557wt_1401

HP E9340A (faulty) for reverse engineering
http://www.ebay.co.uk/itm/290761501523?ssPageName=STRK:MESINDXX:IT&_trksid=p3984.m1436.l2649#ht_500wt_1413

HP/Agilent Autoranging MM £200 - Buy it now
http://www.ebay.co.uk/itm/380458242687?ssPageName=STRK:MESINDXX:IT&_trksid=p3984.m1436.l2649#ht_1299wt_545



HP/Agilent 3466A MM £130 - Buy it now
http://www.ebay.co.uk/itm/380459165338?ssPageName=STRK:MESINDXX:IT&_trksid=p3984.m1436.l2649#ht_920wt_1401

Fluke 8050 MM £30 in the US
http://www.ebay.co.uk/itm/130745882653?ssPageName=STRK:MESINDXX:IT&_trksid=p3984.m1436.l2649#ht_2524wt_1401

These are probably better posted (or duplicated) on the buy/sell thread, but since it's part of the discussion I'll add them here for comment and if anybody wants to add them to the other forum feel free.

Those more experienced feel free to post your recommendations for instruments (that might not be obvious) as I am sure it would help a good number of people out, it also gives an idea of the sort of stuff I am looking at. (I have left out a few I am interested in)

[Mechatrommer]

On the pulser, I'm not confident enough yet to make a good pulser, although I am aware of a few circuits out there, I also have bought a couple of comparators, the LT1721, and a Maxim device that I manage to get eventually through Farnell. Above 50MHz things get much tougher, so I want to learn more before delving into the design of higher frequency stuff, I appreciate it's really challenging some of the engineering at high frequencies. The pulser is purely to measure the rise time of a circuit (or components for suitability) I am aware our current scope limits the performance of a good pulser significantly.

get a good and high bandwidth oscilloscope or spectrum analyzer first before anything else if you plan to design anything speed. otherwise you are playing witchcraft like i am with my DS1052E.

[JuiceKing]

For your pulse generator, are you interested mostly in rise/fall time or pulse rate? If you can make do with slower pulse rates (max 1Mhz) then an old HP 215A could be a good choice--it has rise time of around 1nsec, and a pulse rate that can be controlled with the built-in oscillator or with an external trigger. They are inexpensive dinosaurs, but not as common on eBay as the 8112A. The 8112A goes up to 50Mhz and has a lot of control options, but its fastest rise time is around 5nsec.

[T4P]

It's like trying to go faster with a Golf GTI engine, some people place a too big turbo on a too small car,
just like trying to view 433MHz RF with a 100MHz(hacked) DS1052e (i read your website shafri...)

[nctnico]

Not really. Just get an ADF4351 based frequency generator board. It will output square waves with harmonics into the GHz regions.

a

It's a board!
A real Signal Generator is a lot more than that!

A signal generator often does a lot of different things which make it expensive. In many cases you don't need all the extra bells and whisles. A small board for 1 specific purpose is much cheaper.

[maiakaat]

@mechatrommer

We do want to upgrade to an Agilent MSO (although it will probably be at the lower end of the 3000's and it may take some time to come to terms with the price even then, thanks for the advice, the upgradeability offers a better investment over the 2000X series, and they seem to use much more advanced sampling techniques compared with our Rigol - if I have interpreted some of the articles I have read correctly)

If we can afford it it will likely be 6 to 12 months away, and push the money for a sig-gen right down.

[maiakaat]

@juiceking, that's awesome info, thank you, it is purely rise time measurements that is needed (although the oscilloscope limits this to a simple pass or fail given its bandwidth limitations)

@nctnico, building stuff is something I'd like to do, I just want to develop confidence with easier stuff for now (I want to understand the circuits I/we build) but it could definitely work out cheaper.

[KTP]

You just need a fast rising edge pulse?

Put your favorite microcontroller or even a 555 on a breadboard, then solder up a tinylogic 3 buffer chip like the NC7NZ34 (about $0.20 each) deadbug style on a little bit of copper.  Presto! 700ps risetime 4V to 5V pulses into 50ohms.

Here is what it looked like on a X3024A:

[Mechatrommer]

my point is, gather all your money to get a good dso (like KTP's MSO-X-3024A above, my 2cnts recommendation is 4-10X the frequency you want to design, ie 100MHz circuit requires 400MHz scope or higher) forgetting anything else, so you can shorten the afford time from 12 months to say 6 months? ok fair you already got a rigol dso? and desperate need for FG? right, just get the Hantek DDS3x25 and start playing blairwitch right away, at least you'll learn many magic spells during the 12 months, but beware of your hair, they either grow white, or not grow at all later on. i cant recommend a standalone FG that can go 50-100MHz at your price point, someone may have had/will gave you a suggestion. (ps: i wish i have the money)

[maiakaat]

We decided a few months ago to upgrade the DSO once we can justify the need for the scope (ie once at least once of us - me - has learnt enough)

The cost of a fully features Agilent 3000X series means well have to use the upgrade feature and pick which features we can afford (unless they are a bit cheaper in 6 months or so)