|
Homebrew Equipment A place to show, design and discuss the weird and wonderful electronic creations from the hands of individual members. |
|
Thread Tools |
5th Sep 2017, 8:40 am | #61 |
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,867
|
Re: Spectrum Analyser
They don't have to be round. Tin snips will easily create little square ones.
I build bird's nest structure just above the copperclad, using decoupling capacitors as supports. If stuck for a support, it's worth having a bandolier of high value resistors near the bench. This technique works with SMT parts if you have a microscope, a steady hand and use single strands of wire taken from a stripped length of stranded stuff. David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |
6th Sep 2017, 9:50 am | #62 |
Dekatron
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
|
Re: Spectrum Analyser
I think most of the SMD parts I've tried to do with that ended up in the carpet and up the hoover!
|
8th Sep 2017, 8:54 pm | #63 |
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,867
|
Re: Spectrum Analyser
A few SMT parts and a pair of tweezers and you can have endless fun playing microtiddleywinks.
If you use a few 'PostIt' sheets turned upside down, the sticky strip is a good place to put chip resistors etc and not have them scatter. You can even write values with arrows pointing to the parts. Squeeze tweezers together and rub/drag the ends backwards on a piece of wood. Aim to bend them just a little and then bend them back. Turn them and repeat. This sounds bizarre but it matches the tips so that they agree on where to meet. I work with this stuff for a living, so I had to develop ways of not having to chase the things all over the scenery. That done, they're quite manageable. David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |
9th Sep 2017, 2:17 pm | #64 |
Nonode
Join Date: Mar 2013
Location: Leicester, UK
Posts: 2,535
|
Re: Spectrum Analyser
Yes, I played tiddlywinks with a pair of tweezers and SMD transistors a while back. The post-it note idea as well as the tweezer treatment sound like a good tips. I will have to try this next time I have to deal with the little blighters.
|
26th Oct 2017, 2:50 pm | #65 |
Dekatron
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
|
Re: Spectrum Analyser
Resurrecting my attempt at this project again. I managed to get hold of a helical filter finally that appears to work via ebay and an RF signal generator to replace my dead Marconi 2019A also via ebay. It's only a nasty Chinese generic rebrand generator but it should do the job for this. I redesigned the sweep generator without any ICs because I was bored while I was waiting for the other stuff to arrive. It now has an "SCR/PUT emulator" generating the ramp and a current mirror as per: https://wiki.analog.com/university/c...1k/alm-lab-scr
This is fed into a differential amplifier with gain control for sweep width and then another differential amplifier with summing inputs to control start and end frequency. I've actually got reasonable temperature stability out of this by gluing the transistors together and sticking a bit of heat shrink around them. I started with just 2n2222a and 2n2907a's which I have big bags of but moved to plastic 2n3904/3906 parts because the cases aren't connected to a pin which makes life easier! Whole thing is a right mess of transistors sticking out all over the place at the moment as it's all dead bug. I'm going to make a PCB for it when I'm completely happy with it. Can't promise I'll get much further this year as I suddenly ended up with two dead Tektronix oscilloscopes to play with. |
26th Oct 2017, 9:02 pm | #66 | |
Octode
Join Date: Jan 2014
Location: Oban, Scotland, UK.
Posts: 1,129
|
Re: Spectrum Analyser
Quote:
|
|
27th Oct 2017, 1:14 am | #67 |
Dekatron
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
|
Re: Spectrum Analyser
Fortunately this one is relatively low speed so fingers crossed I won’t ruin it
|
27th Oct 2017, 6:37 am | #68 |
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,867
|
Re: Spectrum Analyser
The layout style of a circuit and the care taken in high-frequency decoupling, grounding, etc, isn't only a matter of the frequency range you intend to use it over, it also needs to be good enough across the frequency range your devices have any gain... otherwise it could hoot.
If the thingy is going to be manufactured, then you also need some added margin as well so that your design will survive future improvements to the devices used. All those statements: "The Xyzzy semiconductor corporation has a policy of continued product improvement and specifications may be changed without notice" are really dire warnings! David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |
27th Oct 2017, 9:52 am | #69 |
Dekatron
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
|
Re: Spectrum Analyser
Plenty of decoupling per stage and there is no phase shift or frequency dependent function on the feedback loops in it so it's universally stable in theory. I'm using relatively high fT transistors (2n3904/3906 @ 300MHz) and there's no oscillation or instability at all. Have had a good poke around with my new scope as well which has 200Mhz bandwidth.
It's amazing how easy it is getting something to oscillate. I had a simple common emitter amp on a breadboard the other day and it was kicking off around 170MHz. Turned out that one leg of the decoupling capacitor had broken so there was an accidental oscillator with the stray inductance of the 4mm banana leads to the power supply. It's quite good fun finding these It's the old saying: amplifiers oscillate, oscillators don't! |
27th Oct 2017, 11:34 am | #70 | |
Hexode
Join Date: Jun 2005
Location: Aberystwyth, Wales, UK.
Posts: 358
|
Re: Spectrum Analyser
Quote:
Dave GW7ONS |
|
27th Oct 2017, 11:42 am | #71 |
Nonode
Join Date: Mar 2010
Location: Dukinfield, Cheshire, UK.
Posts: 2,037
|
Re: Spectrum Analyser
I was building something-or-other a while ago and I ended up with an unused *but still powered* emitter follower onboard. It had a length of RG174 coax on its emitter. I'm sure we can all see what's comng!
The project was showing some nasty behaviour and some strange noises-off. And I noticed a large spike on the analyser around 100 MHz. It was remarkably stable. It turned out that the emitter follower was happily honking to itself at a frequency determined by the length of the coax.
__________________
Andy G1HBE. |
27th Oct 2017, 1:11 pm | #72 |
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,867
|
Re: Spectrum Analyser
The "Emitter-follower" oscillator is a well known circuit. Most are unintentional, but there are deliberate ones. It's one of the most popular ones in the 200-1000MHz region, common in sig gens.
David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |
27th Oct 2017, 2:00 pm | #73 |
Dekatron
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
|
Re: Spectrum Analyser
Even jellybean transistors can oscillate at frequencies way above what you might imagine. When I was looking at the AM modulator circuit a while back I used jellybean BC547B transistors and I tried measuring one of the BC547B transistors in isolation on a VNA to see how high in frequency it might be able to oscillate in theory. The VNA data was exported to a simulator and this predicted the BC547B could oscillate at over 600MHz if given the right L and C parts around it and a reasonably high Vce voltage. i.e. even at 600MHz, the model predicts it could generate sufficient negative resistance to oscillate.
__________________
Regards, Jeremy G0HZU |
27th Oct 2017, 5:08 pm | #74 |
Moderator
Join Date: Mar 2012
Location: Fife, Scotland, UK.
Posts: 22,867
|
Re: Spectrum Analyser
As a general guide, you don't want low impedance at RF/VHF/UHF on any more than one leg of a transistor at once.
David
__________________
Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done |
28th Oct 2017, 10:20 pm | #75 |
Dekatron
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
|
Re: Spectrum Analyser
I did a stupid thing. I managed to screw up the design of this incredibly. Well done me. The sweep vs frequency is the wrong way round. Also the linearity of the sweep was terrible due to the circuit loading changing with voltage. This is a combination of rounding at the extents of the sweep and the inability to integrate any kind of network to fix the varactor non-linearity resulted in scrapping it as well. Straight in the bin and back to the drawing board.
Ergo I sat down this afternoon with Tektronix sweep circuits concepts book [1] and did some thorough research on Miller Integrators and came up with the following which actually works nicely when loaded and is easy enough to gate plus it's dead simple! I'm going attach the existing sweep scale and offset circuit and then a diode clamp network to trim the linearity. You can then disconnect the input stage, apply fixed voltages from a precision power supply and trim each for stable voltage vs frequency. Looking at the linearity required, a total of 8 networks should be required which isn't terrible. Also got to play with my freshly restored Tektronix 475 so a happy evening Also found an excuse to buy a pulse generator! [1] http://w140.com/tekwiki/images/a/a4/062-1098-00.pdf |
22nd Dec 2017, 3:04 pm | #76 |
Dekatron
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
|
Re: Spectrum Analyser
Managed to get the sweep generator working with linearity control finally. The thing uses 8x diode clamps driven from a resistive divider each (each one a trimmer). This is followed by an emitter follower. I now have reasonable linearity. I haven't drawn up a schematic for it yet.
I have however got my hands on a 145MHz helical filter finally. Some progress at least. At the current rate I will complete this some time in 2025... |
24th Dec 2017, 10:29 am | #77 |
Nonode
Join Date: Mar 2010
Location: Dukinfield, Cheshire, UK.
Posts: 2,037
|
Re: Spectrum Analyser
No you won't! In my experience, these things are never quite finished as there is always something else to dangle in, from extra IF filters to better-stabilised supplies and on-screen markers. You'll get tired of taking the lid on and off and it will remain off to make life easier.
I built a number of SpecAns back in the 90's & early 2000's, each one better than the last, but I don't think I ever actually accepted any of them as 'finished'! Good luck with it anyway, it sounds like a good 'un.
__________________
Andy G1HBE. |
24th Dec 2017, 11:34 pm | #78 |
Nonode
Join Date: May 2007
Location: Redruth, Cornwall, UK.
Posts: 2,573
|
Re: Spectrum Analyser
Hi.
This thread has inspired me to revisit the Simple Spectrum Analyser (G4PMK) circuit that I built back in 2007 (can't believe it was that long ago). I originally bought the kits of parts from JAB Electronics (JABDOG) but made my own PCBs, and completed the unit, but never got around to testing it as there was an impending house move so it was packed away in storage. I really like the idea of having the spectrum analyser self contained with it's own display that Dave (ParcGwyn) has built. I'd be most interested to see the circuit details for the display unit. I have a half decent oscilloscope tube from a redundant Heathkit scope plus the power transformer etc so this could be achieved quite economically I think. Dave, would you mind sharing the circuit of your set up to give me a few ideas how to go about the addition of the display? Regards Symon. |
24th Dec 2017, 11:52 pm | #79 |
Dekatron
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
|
Re: Spectrum Analyser
Worth looking at some of the 1970s designs in practical wireless. There was one in 1974 which was rather minimal with the CRT bias and ran off DC so marginally less dangerous than others. . Also Heathkit early transistor scopes with the +/-9v rails are pretty easy to follow designs. IO-102 etc.
In reply to Andrew2, nothing I’ve built ever gets to the “lid” stage |
25th Dec 2017, 12:16 am | #80 |
Nonode
Join Date: May 2007
Location: Redruth, Cornwall, UK.
Posts: 2,573
|
Re: Spectrum Analyser
Thanks for the tips Mr Bungle, I'll have a good look through the PW circuits.
Regards Symon. |