Author Topic: The perfect function generator, does it exist ?  (Read 5302 times)

0 Members and 1 Guest are viewing this topic.

Offline Kiriakos-GR

  • Super Contributor
  • ***
  • Banned!
  • Posts: 3525
  • Country: gr
  • User is banned.
    • Honda AX-1 rebuild
The perfect function generator, does it exist ?
« on: December 11, 2011, 07:17:06 am »
I did face an issue with an professional but old function generator.
At at higher frequencies the named voltage output ( Peak to Peak) above the 300KHz it starts to drop significantly,
more specifically the generator have as specs 20VPP 0-20MHz,  and at 300KHz the voltage drops to 5VPP.
And above 300KHz  it drops even more, to the point the output to become few mV.

Is this normal or expected ?

In theory I would  expect 20VPP in all the range of frequencies !!
Is out there the perfect Function generator, that has identical output in full range ?   

My small own small and cheap frequency generator 27KHz max, acts the same, but so far I was thinking that it was due the simplified design. 
 

Offline Zad

  • Super Contributor
  • ***
  • Posts: 1013
  • Country: gb
    • Digital Wizardry, Analogue Alchemy, Software Sorcery
Re: The perfect function generator, does it exist ?
« Reply #1 on: December 11, 2011, 08:04:49 am »
I think you would have to define "perfect".

20vpp equates to 8W into a Zo=50R, That is quite a lot of power, and a heck of a lot to ask from a signal generator.

Does it state what the characteristic impedance is, or what the test load impedance is?

Offline Kiriakos-GR

  • Super Contributor
  • ***
  • Banned!
  • Posts: 3525
  • Country: gr
  • User is banned.
    • Honda AX-1 rebuild
Re: The perfect function generator, does it exist ?
« Reply #2 on: December 11, 2011, 08:17:40 am »
The impedance is adjustable to 50 or 600 ohm.

Even so I need 2.5VPP up to 2MHz, do I ask too much ?   
 

Offline Conrad Hoffman

  • Super Contributor
  • ***
  • Posts: 1057
  • Country: us
    • The Messy Basement
Re: The perfect function generator, does it exist ?
« Reply #3 on: December 11, 2011, 08:21:04 am »
There's either something wrong with it, or your measurement, or you've got it loaded too much. Even though a generator may have a 50 ohm output impedance, it's rare that they can drive 50 ohms to any large amplitude. It's important to understand the difference. It should be quite flat into a high impedance load over the full frequency range.

FWIW, the perfect function generator is the long gone Wavetek 185 sweep generator. It only goes to 5 MHz, but has separate dials for start and stop frequency. They had others with more bandwidth, but above a couple MHz I just switch to a real RF generator.
 

Offline Richard W.

  • Regular Contributor
  • *
  • Posts: 149
  • Country: de
Re: The perfect function generator, does it exist ?
« Reply #4 on: December 11, 2011, 08:31:12 am »
Hello Kiriakos,

do you speak about sine wave or rectangular? du you measure the voltage with a scope or multimeter or even a hf-voltmeter?
Maybe your readings are caused by the frequency response of your measuring instrument. The output voltage of my HP 8111A is stable up to it's maximum frequency.

Quote from: Zad
20vpp equates to 8W into a Zo=50R, That is quite a lot of power, and a heck of a lot to ask from a signal generator.

it's not this much power. You have to use RMS voltage for P=U²/R
« Last Edit: December 11, 2011, 08:36:42 am by Richard W. »
 

Offline Zad

  • Super Contributor
  • ***
  • Posts: 1013
  • Country: gb
    • Digital Wizardry, Analogue Alchemy, Software Sorcery
Re: The perfect function generator, does it exist ?
« Reply #5 on: December 11, 2011, 08:34:34 am »
1/8W is not unreasonable (2.5V into 50R)

If you have any decent op-amps in stock, maybe consider hacking together a simple buffer?

(Re:8W, yep my mishtake, 4W for a sinusoidal)
« Last Edit: December 11, 2011, 08:40:52 am by Zad »
 

Offline ejeffrey

  • Super Contributor
  • ***
  • Posts: 1818
  • Country: us
Re: The perfect function generator, does it exist ?
« Reply #6 on: December 11, 2011, 09:33:29 am »
At small signal levels (say 1 Vrms) into 50 ohm any function generator should be able to deliver the rated bandwidth within 3 dB or better.

There are plenty of pitfalls but nothing I can think of that would account for the signal dropping to the millivolt levels.
 

alm

  • Guest
Re: The perfect function generator, does it exist ?
« Reply #7 on: December 11, 2011, 09:38:54 am »
There's either something wrong with it, or your measurement, or you've got it loaded too much. Even though a generator may have a 50 ohm output impedance, it's rare that they can drive 50 ohms to any large amplitude.
The amplitude should of course be reduced by 50% with a 50 ohm load, but apart from that it should be quite capable of reaching its max amplitude, otherwise it doesn't have a true 50 ohm output.

It should be quite flat into a high impedance load over the full frequency range.
The good ones are, the cheap ones less so, but millivolts at its max frequency sounds very low. Philips should be decent, although not in the same league as eg. HP.

FWIW, the perfect function generator is the long gone Wavetek 185 sweep generator. It only goes to 5 MHz, but has separate dials for start and stop frequency. They had others with more bandwidth, but above a couple MHz I just switch to a real RF generator.
Most RF generators don't seem to do square waves and have a fairly low amplitude. They also seem to have more issues with unterminated loads.

(Re:8W, yep my mishtake, 4W for a sinusoidal)
4W? I estimate about 1W. 4W would be cooking those terminators which are often rated for only 1/2W or 1W.
 

Offline Kiriakos-GR

  • Super Contributor
  • ***
  • Banned!
  • Posts: 3525
  • Country: gr
  • User is banned.
    • Honda AX-1 rebuild
Re: The perfect function generator, does it exist ?
« Reply #8 on: December 11, 2011, 11:34:19 am »
Hello Kiriakos,

do you speak about sine wave or rectangular? du you measure the voltage with a scope or multimeter or even a hf-voltmeter?
Maybe your readings are caused by the frequency response of your measuring instrument. The output voltage of my HP 8111A is stable up to it's maximum frequency.

I use wave and I am testing the max true bandwidth of four multimeters under test. 
The leader is this race is the Brymen BM869.

Is any rule of the thump about from what frequency and over its best to use the 50Ohm terminator ?
I bet that a 600Ohm terminator would cause less voltage drop , but I do not have one.
« Last Edit: December 11, 2011, 11:37:06 am by Kiriakos-GR »
 

Offline Mechatrommer

  • Super Contributor
  • ***
  • Posts: 8741
  • Country: my
  • reassessing directives...
Re: The perfect function generator, does it exist ?
« Reply #9 on: December 11, 2011, 11:43:03 am »
build a peak detector and measure with the dmm vdc.
if something can select, how cant it be intelligent? if something is intelligent, how cant it exist?
 

Offline ejeffrey

  • Super Contributor
  • ***
  • Posts: 1818
  • Country: us
Re: The perfect function generator, does it exist ?
« Reply #10 on: December 11, 2011, 08:35:06 pm »
You are seriously measuring the output with a multimeter with 100 kHz bandwidth and surprised that the amplitude drops at 300 kHz?  Your function generator is fine, you have a serious case of user error.
 

Offline Kiriakos-GR

  • Super Contributor
  • ***
  • Banned!
  • Posts: 3525
  • Country: gr
  • User is banned.
    • Honda AX-1 rebuild
Re: The perfect function generator, does it exist ?
« Reply #11 on: December 11, 2011, 10:59:01 pm »
Your function generator is fine, you have a serious case of user error.

Could be.
This is why I am now with the calibration manuals at hand.
So to avoid testing at extreme conditions not capable multimeters.
 

Offline Richard W.

  • Regular Contributor
  • *
  • Posts: 149
  • Country: de
Re: The perfect function generator, does it exist ?
« Reply #12 on: December 11, 2011, 11:12:32 pm »
Quote from: Kiriakos-GR
I use wave and I am testing the max true bandwidth of four multimeters under test.
The leader is this race is the Brymen BM869.
The Brymen's bandwidth is "only" 100kHz. So, what do you expect?
That's fine if it's measuring accurate up to 300kHz.

Quote from: Kiriakos-GR
Is any rule of the thump about from what frequency and over its best to use the 50Ohm terminator ?
I bet that a 600Ohm terminator would cause less voltage drop , but I do not have one.
50 ohm output -> 50 ohm terminator
600 ohm output -> 600 ohm terminator

If you want to measure the hf voltage of your generator you need either an oscilloscope or a hf-probe (peak detector).
 

Offline Kiriakos-GR

  • Super Contributor
  • ***
  • Banned!
  • Posts: 3525
  • Country: gr
  • User is banned.
    • Honda AX-1 rebuild
Re: The perfect function generator, does it exist ?
« Reply #13 on: December 11, 2011, 11:15:17 pm »
Other than AC bandwidth I am testing also the frequency counter, thanks for the tips.  :)
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 12459
  • Country: gb
  • Hero999
Re: The perfect function generator, does it exist ?
« Reply #14 on: December 12, 2011, 03:12:56 am »
Perhaps you should look at the output of the signal generator on an oscilloscope with 10 times tha bandwidth of the squarewave, then you can be sure it's the signal generator playing up and not the meter.
 

Offline Kiriakos-GR

  • Super Contributor
  • ***
  • Banned!
  • Posts: 3525
  • Country: gr
  • User is banned.
    • Honda AX-1 rebuild
Re: The perfect function generator, does it exist ?
« Reply #15 on: December 12, 2011, 05:11:08 am »
Well at the end, I did run the tests that I had in mind by the book of the manufacturers, about verifying voltage and Duty cycle accuracy.

And I got wild only by using the multimeters under test, as frequency counters.
The results about the max frequency displayed with shine wave and square wave differs allot,
and this comparison was impressive and allot of fun about doing it.
The 0-20MHz Function generator did helped allot.  ;)
« Last Edit: December 12, 2011, 05:12:56 am by Kiriakos-GR »
 

Offline Richard W.

  • Regular Contributor
  • *
  • Posts: 149
  • Country: de
Re: The perfect function generator, does it exist ?
« Reply #16 on: December 12, 2011, 07:03:17 am »
Thats the frequency response of my HP8111A, measured with a Rohde&Schwarz URV4 Millivoltmeter + URV-Z2 insertion Unit.
I set the function generator to 5V and changed the frequency from 10kHz up to 22MHz.
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 12459
  • Country: gb
  • Hero999
Re: The perfect function generator, does it exist ?
« Reply #17 on: December 12, 2011, 08:46:48 am »
The voltage recorded by the meter will decline past 10kHz or so, then it will increase due to ringing, reflections etc. regardless of the meter's quality ,unless the probes are frequency compensated which isn't normally the case.
 

Offline Kiriakos-GR

  • Super Contributor
  • ***
  • Banned!
  • Posts: 3525
  • Country: gr
  • User is banned.
    • Honda AX-1 rebuild
Re: The perfect function generator, does it exist ?
« Reply #18 on: December 12, 2011, 11:01:29 am »
As cables I had use coaxial RG-58 by CAVEL (Made in Italy) the highest quality ever, plus BNC connectors, BNC bridges and in some cases even terminators 50ohm.

The specific function generator its not my, so to care about complete calibration for it.
I found errors on the frequency counter / Offset voltage/ max output of duty cycle is less than the specified.
Its a nice device, but as analogue one, its easy to use, but hard to set an very specific setting as voltage or frequency. 
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf