UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > Specific Vintage Equipment > Vintage Test Gear and Workshop Equipment

Notices

Vintage Test Gear and Workshop Equipment For discussions about vintage test gear and workshop equipment such as coil winders.

Closed Thread
 
Thread Tools
Old 29th Oct 2018, 5:58 am   #21
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: RF power meter calibration

If you're rebuilding a dummy load, it may be worth adding an attenuated output if it's mechanically feasible.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now  
Old 29th Oct 2018, 1:10 pm   #22
MrBungle
Dekatron
 
MrBungle's Avatar
 
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
Default Re: RF power meter calibration

As requested by another forum member, I have posted my power meter implementation here: https://vintage-radio.net/forum/show....php?p=1087121

This is good for low level signals only. It needs a lot of attenuation to measure power at a higher level. Based on W7ZOI design.
MrBungle is offline  
Old 29th Oct 2018, 6:38 pm   #23
Bazz4CQJ
Dekatron
 
Bazz4CQJ's Avatar
 
Join Date: Apr 2009
Location: Oxfordshire, UK.
Posts: 4,923
Default Re: RF power meter calibration

Quote:
Originally Posted by Radio Wrangler View Post
If you're rebuilding a dummy load, it may be worth adding an attenuated output if it's mechanically feasible.David
The dummy load referred to in previous posts is shown as attached. The updated plan for it is to replace that existing flange resistor with a comparable item uprated from 150 to 250W. My current interests (having now tried and totally abandoned 2m) is try out 5MHz with my IC-718, which should be pushing out ~100W.

Any suggestions about how to incorporate an attenuated output would be welcome. I'd not thought about it previously, so will need to do a little research. Let's think of this as primarily for use at 30MHz and below.

I think that your MkII does indeed look very neat Mr Bungle; I shall have to try get the W7ZOI nearer to the top of my to-do list.

B
Attached Thumbnails
Click image for larger version

Name:	P1020342.jpg
Views:	110
Size:	100.9 KB
ID:	171787  
__________________
Saturn V had 6 million pounds of fuel. It would take thirty thousand strong men to lift it an inch.
Bazz4CQJ is offline  
Old 29th Oct 2018, 7:20 pm   #24
Bazz4CQJ
Dekatron
 
Bazz4CQJ's Avatar
 
Join Date: Apr 2009
Location: Oxfordshire, UK.
Posts: 4,923
Default Re: RF power meter calibration

This article on RF power measurement http://www.ab4oj.com/test/pwrmeas.html seems informative and pleasingly basic. Please tell me if I'm wrong.

B
__________________
Saturn V had 6 million pounds of fuel. It would take thirty thousand strong men to lift it an inch.
Bazz4CQJ is offline  
Old 29th Oct 2018, 7:34 pm   #25
MrBungle
Dekatron
 
MrBungle's Avatar
 
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
Default Re: RF power meter calibration

Yes all of those methods do the job to be honest. That's what I was doing before I built a power meter.

The reason I built the power meter though is that it has a log response vs power in which is useful for watching filter roll off
MrBungle is offline  
Old 29th Oct 2018, 8:40 pm   #26
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: RF power meter calibration

A suitable amount of attenuation would be around 30dB

The standard figures for a 30dB pi section attenuator at 50 Ohms is 53.3 Ohms shunt, 790 Ohms series, 53.3 Ohms shunt.

A 3.3 Ohms RF resistor able to handle 6% of the input power could be used to pad out the dummy load 50 Ohm resistor.

Something similar but scaled down in power by a factor of 1000 could go on the attenuator output as the shunt element.

The series element of 790 Ohms needs to handle about 6.7% of the input power. It could be broken up into series combination of resistors to split up the power and the voltage.

I've used lowish inductance metal film resistors successfully at HF.

A screen to stop the output loop seeing mag field from the input loop and to shield electric fields is a good idea.

Getting good performance at HF is easy enough, getting the thing to be flat to VHF so you can measure harmonics is a lot harder.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now  
Old 30th Oct 2018, 12:52 am   #27
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: RF power meter calibration

Quote:
Originally Posted by Bazz4CQJ View Post
This article on RF power measurement http://www.ab4oj.com/test/pwrmeas.html seems informative and pleasingly basic. Please tell me if I'm wrong.

B
I only skimmed the article but I don't see how the divider in fig 2 will work across the HF band with a typical 10:1 scope probe because of the capacitance of the scope probe. I suppose the 9k1 resistor could be compensated with a low value capacitor across it but the article doesn't mention this. If the 9k1 resistor was compensated to suit the scope probe then the compensation would then be wrong if was then used with a diode probe instead.

The heatsink on your dummy load does look a bit small (for 100W) and I'm surprised it even manages the 1degC/W your earlier graph suggests. It's probably OK for short bursts or for 100W PEP SSB with a typical speech waveform but even then it is going to get quite hot I think.
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 30th Oct 2018 at 12:59 am.
G0HZU_JMR is offline  
Old 30th Oct 2018, 4:07 am   #28
Bazz4CQJ
Dekatron
 
Bazz4CQJ's Avatar
 
Join Date: Apr 2009
Location: Oxfordshire, UK.
Posts: 4,923
Default Re: RF power meter calibration

Well, doing a bit more research, it's clear we are going down a beaten track https://www.vintage-radio.net/forum/...ad.php?t=24608.. sadly .

B
__________________
Saturn V had 6 million pounds of fuel. It would take thirty thousand strong men to lift it an inch.
Bazz4CQJ is offline  
Old 30th Oct 2018, 8:40 am   #29
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: RF power meter calibration

It's a good thread, but the photo oh the 1740A scope shows a very bad setup. The 1740 has a switchable 50 ohm resistor you can put in circuit on its input. I think it's OK to 0.125W. So seeing a T-piece connecting the scope straight to the transmitter output in parallel with a much beefier dummy load.

This is a recipe for burning out the scope's 50 ohm terminator if you switch it in.

Correct usage would be to not use a plain dummy load, but to use a high power attenuator to reduce the power to the scope to around 100mW, then switch in the scope input to 50 ohms, and the scope will now correctly terminate the coax feed to its input. AND the attenuator will see the load impedance it is designed to work with.

The other photo shows the rather cute 200MHz Tek 2022 scope. I use these daily at work. They don't have the 50 ohm termination switch. With these you have to add an external T-piece and a small 50 ohm termination, right onto the front panel connection. Better yet, I have a few BNC through-connection terminations for this job.

It's a smart move by Tek to not have a 50 ohm terminator in a basic model scope given people's track record of burning the things out. Having one, like in the HP1740, is a great convenience if no-one has already burned it out You learn not to lend out these scopes.

In that thread there is also a request for a home brew directional coupler. I put one in Sprat about 20 years earlier... "A dual-directional power meter" Sprat 61 and all the G-QRP club Sprat compilation CD roms. The design contains a directional coupler good for up to 200W with a pair of 50 ohm terminations and a pair of diode detectors. Dump the terminations and detectors and you have a good HF directional coupler. Flat to 0.1dB across the 1.8-30MHz range. BUT remember that both coupled outputs must always see good 50 ohm loading.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now  
Old 30th Oct 2018, 7:35 pm   #30
Bazz4CQJ
Dekatron
 
Bazz4CQJ's Avatar
 
Join Date: Apr 2009
Location: Oxfordshire, UK.
Posts: 4,923
Default Re: RF power meter calibration

I stumbled upon a comment on this subject made on an American forum, suggesting that outside of a well-equipped laboratory, using only the sort of equipment an average field engineer or amateur was likely to have, errors of 10-20% could be experienced at even as low as 30MHz.

I don't feel qualified to comment on that except to say that it does seem to sum up some of the comments made in the two threads on this forum.

I think the one thing to come out of this thread for me is that any attempt to do RF power measurement starts with having a good dummy load, and I'm now sure that the one I've been playing with is inadequate for 100W. So, I will await the arrival of the mule train with the 250W aluminium nitride resistors, which can go on one of the much heftier heat sinks in my junk box. I think I'll add a thermocouple at one end, and an output attenuator at the other and 'compare and contrast' the two.

B
__________________
Saturn V had 6 million pounds of fuel. It would take thirty thousand strong men to lift it an inch.
Bazz4CQJ is offline  
Old 30th Oct 2018, 11:10 pm   #31
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: RF power meter calibration

Just to recap, it's possible to predict how poorly the divider below will work with a x10 scope probe because a decent x10 scope probe will have about 12pF of internal capacitance. So there will be an RC filtering effect with an effective R of about 900R and C of 12pF. So the divider + probe would probably roll off by about 3dB by 14MHz... hopeless.

I think the classic scope method is going to be adequate if someone just wanted to get an idea of their transmit power but I'm not sure it is the right tool to use for checking and adjusting the calibration of a fairly decent ham power meter.

There are lots of traps for the unwary even if trying for a fairly casual measurement. The picture with the HP1740 scope in the link shows far too much cable in the tap line to the scope and this will have a lot of capacitance when the scope is set to the 1Meg input impedance setting. It could even look like a stub filter at higher frequencies so it will load the transmitter quite a bit at higher frequencies. Obviously, the scope's internal 50R termination can't be used here or it would be damaged instantly.

It's also possible to stress or even damage the scope even when set to the 1meg input. This is because the picture shows a direct connection to the scope. This is probably OK at very low RF frequencies but a typical old school analogue scope will not be designed to cope with high voltages at RF frequencies up towards 28MHz. By 70MHz or 145MHz I think it would be possible to damage a typical analogue scope with a fairly low power transmitter.

The same applies to a lot of modern digital scopes. Usually it's best to derate the voltage by a factor of two for every doubling of frequency so once well up into VHF the sensible/safe level could just be a few Vrms in both 1meg or 50R input settings.

If you use an external 50R resistor and leave the scope in the 1Meg setting you can expect to see the internal capacitance of the scope begin to affect the VSWR by 28MHz and also the scope itself won't look as simple as 1Meg in parallel with 20pF by 28MHz. This is partly what makes it vulnerable to high voltages at high frequencies and it will also (very slightly) alter the resistance of the load away from 50R up at 28MHz. Also, a typical 75MHz scope will be losing some accuracy by 28MHz. So lots of factors creep in that affect accuracy and reliability as you go up in frequency.
Attached Thumbnails
Click image for larger version

Name:	Vdivider.gif
Views:	88
Size:	12.9 KB
ID:	171853  
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 30th Oct 2018 at 11:20 pm.
G0HZU_JMR is offline  
Old 30th Oct 2018, 11:42 pm   #32
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: RF power meter calibration

There was a design for a self-balancing bolometric power meter in Radcom once. It used a pair of small filament bulbs as thermistors. Quite home-brewable for HF and maybe as good as you can get without spending money on a lab type meter.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now  
Old 2nd Nov 2018, 3:26 pm   #33
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: RF power meter calibration

Quote:
I think the one thing to come out of this thread for me is that any attempt to do RF power measurement starts with having a good dummy load
Yes, and I also think it is important to know how good it is if doing something like a power meter calibration check. If using an analogue scope or the diode detector method the resistance of the load (at RF) needs to be known. Buy a decent load rated at several times the wanted frequency and it can often be enough to just check the load at DC with a DMM. But it's best to check it with a VNA if at all possible.

I'm surprised that the diode detector method seems to be getting dismissed or ignored. Before I managed to buy an HP432A power meter I used to check my Anritsu thermocouple power meter using a peak detector and a DVM and a calibrated 30dB attenuator. I did this at 27MHz and at 145MHz using a CB radio and a 2m radio each producing about 3 or 4W. I did this for several years and always got consistent results when comparing the power calculated by the detector vs the power measured by the power meter via the attenuator. I'd like to think it was within 2% but it would definitely be within 5% using this method. Both radios had harmonics below -60dBc.
I used a classic old HP2800 Schottky diode as the detector and a decent SMD ATC cap after the diode. So the detector should be good and flat up into the VHF region.

These days I use the HP432A and HP478A sensor and the DC substitution method to check the Anritsu power meter and its 0dBm 50MHz reference. I've had an HP431C power meter for over 25 years but it wasn't as easy to do the substitution method with this older meter.

I also check for flatness vs frequency using various methods up to several GHz and the most recent method I used at home was with my old HP432A power meter as part of the (error correction) feedback in the setup below. This uses a classic old HP 83752A 20GHz sweep generator and I'm lucky to own one of these. A fabulous piece of RF engineering

This method is fast but it isn't as good as using the DC substitution method but it does give very good results over a huge RF bandwidth. If I'd included the correction factor vs frequency for the efficiency of the HP478A sensor the results would have been flatter at the band edges.

But it is a quick/automated method to compare two power meters over frequency to see if they agree and this level of agreement is more than good enough for me

Note that the resistive splitter might initially appear 'wrong' in this setup because it uses a pair of 50R resistors inside. But because there is negative feedback via the power meter and ALC, the centre feed point of the splitter becomes a low impedance voltage source. So to make it a 50R source it needs to be followed by a series 50R resistor at both output ports. By using a precision splitter here it is possible to create an almost perfect 50R source and the amplitude accuracy will be as good as the power meter in the feedback path. One other application for this method is to use it to measure the frequency response of a microwave spectrum analyser. This method ensures good source flatness and also excellent source match and this really helps to minimise the overall measurement uncertainty.
Attached Thumbnails
Click image for larger version

Name:	ALC_setup.jpg
Views:	93
Size:	45.6 KB
ID:	172013   Click image for larger version

Name:	alc_fb2.jpg
Views:	59
Size:	70.4 KB
ID:	172014  
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 2nd Nov 2018 at 3:44 pm.
G0HZU_JMR is offline  
Old 2nd Nov 2018, 6:28 pm   #34
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: RF power meter calibration

By the way, If anyone reasonably nearby needs an RF power meter calibrating in the HF/VHF region, I have a thermal DC-substitution standard and can source up to 200W on HF, 100W on 2m about 40W on 70cm and 10w on 23cm.

David (G-QRP number withheld at licencee's request )
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now  
Old 2nd Nov 2018, 7:33 pm   #35
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: RF power meter calibration

See below for a plot showing the performance of a decent (but cheaply available) RF attenuator rated to 8.5GHz.

For medium power stuff there are some great deals out there for RF attenuators. I would always recommend that the attenuator be rated at several times the frequency and power you actually want to use it at.

Many years ago I bought a few brand new Weinschel attenuators from the states. These were 30dB 50W rated to 8.5GHz with cal certificates and they were only £35 each. Someone in the states had dozens of them for sale and the price was low! You can still get them today from dealers in the UK for about £80 and I reckon these will be unsold stock from the same batch from the states.

See below for the performance when measured on my VNA after an Ecal calibration. The VNA display shows a return loss of about 45dB over 0-150MHz and the attenuation is a flat 30.12dB. These were precision devices and sadly, I sold two to a colleague. I only have this one left. In order to preserve its performance I only ever run it at power levels up to about 12W PEP and only for short bursts.

For higher power stuff I have a huge custom attenuator from the classic Bird WA series. Mine is a special 40dB version and I think it is rated to 1kW up to 2.5GHz. It is very accurate and also has very good return loss.

https://www.birdrf.com/Products/Test...tenuators.aspx

But for 100W on the HF bands you could probably make an accurate 50R dummy load that would have very low VSWR as long as you bought good quality resistors (and arranged the values such that it is very close to 50.0R)and used it for short bursts only.
Attached Thumbnails
Click image for larger version

Name:	30db(wein).gif
Views:	57
Size:	18.3 KB
ID:	172023  
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 2nd Nov 2018 at 7:43 pm.
G0HZU_JMR is offline  
Old 2nd Nov 2018, 8:21 pm   #36
Radio Wrangler
Moderator
 
Radio Wrangler's Avatar
 
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,799
Default Re: RF power meter calibration

While we're mentioning attenuators, it's a good time for a reminder

Usually medium and high power attenuators have a high power end and a low power end. Check which one you're about to apply power into before turning on.

Sadly a lot of people instinctively connect-up whichever way round the sexes of the connectors are most convenient for what they're connecting it to.

Using one the wrong way round can ruin it

Some of the Weinshel medium power ones are deliberately symmetrical and can be used either way round. But they are somewhat in the minority.

Some attenuators are not clearly marked, and some you have to download a datasheet to find out.

This also means that second hand attenuators are sometimes unscrupulously sold on.

David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done
Radio Wrangler is online now  
Old 2nd Nov 2018, 9:41 pm   #37
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: RF power meter calibration

Yes, the big 1kW Bird attenuator is directional and the high power end has a series of graded attenuators starting at about 0.5dB and each section gradually gets bigger until 40dB overall is achieved. This approach keeps the heat evenly distributed along the attenuator.

I dug out the old FM CB radio and my diode detector and ran it up again. I think it's the first time I've done this in several years as my records show I last did this in 2014 with an error of -0.05dB compared to the power meter.


Here's the results for today. This is at 27MHz and the CB connects to the diode detector and uses the Weinschel attenuator as the dummy load. The far end of the attenuator goes to an Anritsu thermocouple power meter with an offset of 30.12dB programmed in to suit the attenuator.


Power Meter Reading 36.03 dBm

DMM Reading 19.60 V
Diode Vdrop = 0.35 V

Corrected for Diode Vdrop 19.95 V


P(det) = 3.98 W
P(det) = 5.999 dBW
P(det) = 35.999 dBm

Weinschell 8GHz Attenuator 30.12 dB


Power Meter Reading 36.030 dBm
P(det) = 35.999 dBm


Error = -0.031 dB

This is very close to the results I got 4 years ago. I also do this at 145MHz with the same diode detector and power meter and attenuator and get similar agreement.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 2nd Nov 2018, 10:16 pm   #38
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: RF power meter calibration

With the CB PSU voltage turned down the power was adjusted to a lower value (was 4W now 1.5W) and I ran the tests again.

Power Meter Reading 32.00 dBm

DMM Reading 12.20 V
Diode Vdrop = 0.35 V

Corrected for Diode Vdrop 12.55 V


P(det) = 1.58 W
P(det) = 1.973 dBW
P(det) = 31.973 dBm



Weinschell 8GHz Attenuator 30.12 dB


Power Meter Reading 32.000 dBm
P(det) = 31.973 dBm


Error = -0.027 dB

The diode type used in this detector is the classic old HP2800 Schottky detector diode as per my old stash of them seen below. This should be very similar to the modern 1N5711 Schottky diode.
Attached Thumbnails
Click image for larger version

Name:	hp2800.jpg
Views:	75
Size:	56.4 KB
ID:	172043  
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 2nd Nov 2018 at 10:35 pm.
G0HZU_JMR is offline  
Old 3rd Nov 2018, 12:15 am   #39
Radio1950
Hexode
 
Join Date: Apr 2017
Location: Buderim, Queensland, Australia.
Posts: 428
Default Re: RF power meter calibration

Quote:
Originally Posted by G0HZU_JMR View Post

But for 100W on the HF bands you could probably make an accurate 50R dummy load that would have very low VSWR as long as you bought good quality resistors (and arranged the values such that it is very close to 50.0R)and used it for short bursts only.
I made up such a HF load using an RFP 250W Termination obtained from eBay for very low cost. I derate the load to 100 W, duty cycle 25%.
The copper sheet is important and the coax is bent deliberately.
This is made as a practical load rather than an accurate termination.
Attached Thumbnails
Click image for larger version

Name:	1   50R 100W  DSC06641.jpg
Views:	90
Size:	35.8 KB
ID:	172045   Click image for larger version

Name:	2   50R 100W  DSC06632.jpg
Views:	103
Size:	40.6 KB
ID:	172046  
Radio1950 is offline  
Old 3rd Nov 2018, 4:03 pm   #40
Bazz4CQJ
Dekatron
 
Bazz4CQJ's Avatar
 
Join Date: Apr 2009
Location: Oxfordshire, UK.
Posts: 4,923
Default Re: RF power meter calibration

Quote:
Originally Posted by Radio1950 View Post
I made up such a HF load using an RFP 250W Termination obtained from eBay for very low cost. I derate the load to 100 W, duty cycle 25%.The copper sheet is important and the coax is bent deliberately.
This is made as a practical load rather than an accurate termination.
Thanks for showing us that. I'm currently waiting for some similar 250W terminations to arrive and will then be building something very similar, again using a copper base mounted on a bigger heat sink. What is the rationale for the bend in the coax? I realise that heat going in to the connector may be an issue.

I'm definitely going to fit a temperature sensor (TC or thermistor) to the chip with the intention of comparing DC inputs with RF inputs. I believe that maximum chip temperature is 100'C, so that needs to be watched too.

B
__________________
Saturn V had 6 million pounds of fuel. It would take thirty thousand strong men to lift it an inch.
Bazz4CQJ is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 5:09 pm.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.