Randys Bench

A repository for projects past and present

Leave a comment

Neon Wavelength Calibrator

Since I’ve finally started back to do some work on a sadly neglected project, the Compact solar spectrograph, I thought it time to put together a proper neon-based spectral (wavelength) calibrator for this instrument.

I researched a few designs for the power supply, including a older Forrest Mim’s engineer’s notebook 555-based design shown below:

As well I took a look at the internal calibrator incorporated in Shelyak’s Lhires III spectrograph design which is the one I decided on. This uses a UC3843 PWM controller which, coming from the power supply industry, I’d used in design’s back in the 90’s and became familiar with. This circuit incorporates both PWM regulation control and voltage multiplier circuitry to provide a DC output of 100V. It appears to work well with a variety of neon bulb types.

Here’s a look at the completed card assembly:

This will be a nice addition to the spectrograph redesign upcoming over the next few months. Here is a nice document describing the general attributes of neon bulbs and a few silly circuits as well…Neon Lamp Information. Below is a wavelength chart of the neon gas spectral signatures to give some idea of the calibration coverage one would expect. There’s also some finer lines in the deep blue and violet end of the spectrum:

And speaking of the spectrograph redesign underway, here’s what it look’s like at this stage of development:

Quite a ways to go yet….back soon





Leave a comment

Objective Prism Spectrography

One might think given the title of this post that there is a fundamental difference between recently presented posts on HF frequency analyzers and that of light frequency decomposition and analysis. In reality it’s really more a matter of frequency range and components needed to analyze the frequency content. Where as in the HF band we’re dealing with frequencies in the sub 100 MHz range, light frequencies occupy a significantly higher frequency segment of the electromagnetic spectrum, typically in the 400 to 750 THz range!

Back in July of 1998 I had the opportunity for a brief correspondence with the author of a particular Sky & Telescope article  published in May of 1983 concerning the hardware arrangement and procedures for capturing stellar spectra using an objective prism. This involves placing a 5, 45 or 60 deg. glass prism at the front of a suitable 35mm camera (or other) and lens. The camera’s shutter is opened and the object’s thin spectrum is allowed to widen using the earth’s diurnal motion (or varying the speed of mount’s RA clock drive). The use of an objective prism has been popular for a very long time due to it’s simplicity, light efficiency and it’s capability to capture a great number of stars in a single exposure. Excellent for sky survey purposes. Although Brent Sorensen’s May article was certainly not the 1st to be published in popular Astronomy magazines concerning this subject matter it (along with other’s)did manage to further my interest and improve the somewhat crude setup I’d had up to that point and proved to be a solid path in realizing an improved DIY system to explore the realm of stellar spectra.
The subject of the correspondence was, not only to express my appreciation for the article’s subject matter, but to ask if further images were captured which may have not been published in the original article. Turns out there were quite a few, and Brent graciously consented to place the negatives on loan to me. I had the negatives made into prints at a local camera shop and later scanned them into my computer.

Sorensen’s setup at Lake Afton’s Public Observatory Wichita, Kansas – Fall 1982

Below is a selection of these never before published prints using 45 deg and 60 deg prisms:

45 deg prism capture of Orion area

45 deg prism capture of M45 (Pleiades)

45 deg prism capture of  the Hyades (Taurus)

60 deg prism capture of Aldebaran (Taurus) 4min exp.

60 deg prism capture of pollux 3min exp.

60 deg prism capture of Sirius 3min exp.

Here’s an additional link for Brent’s followup article describing his reflection grating spectrograph

And I think that’s my story today…Randy


Leave a comment

Wrapping up the MC3356 Spectrum Analyzer

Oop’s! Well this time I managed a self inflicted circuit error. I was thinking I needed to invert the VCO sweep output hence adding an inverting x1 opamp to the VCO output…wrong. Due to the way the 1st mixer’s LO’s varactor is reversed biased the sweep polarity needs to be opposite of that of the X-axis sweep ramp. Works a lot better now. I’ve picked up another 30 MHz or so on the LO’s high frequency range. The occasional pitfalls of over-thinking…I’ve updated the Sweep generator/video amplifier schematic posted earlier. Below is the salient bit:

So what else…I corrected an additional error of judgement by removing the MC3356 IC socket I originally decided to use. By removing said socket I eliminated approx. 4.5mm of lead length and it’s accompanying inductance and parasitic capacitance(s). Makes for a tighter layout for those VHF frequencies involved in the SA design. Additionally I’m changing over the leaded radial caps I’m using around the MC3356 LO for SMT 1206 leadless components, and moving the varactor & L2 coil to a tie point directly to pin 3. I’d like to push the SA’s upper end as high as possible. Here I’ve picked up an additional 5-10MHz of frequency span.

At this point I’ve settled on an L2 coil construction of 2.5 turns of 3/16″ diameter. L2 constuction depends on your particular component layout. The tuning varactor remains a BB209 variety.

Lastly I played around with characterizing a few 10.7 MHz ceramic filters I’ve got in the bins to see if this is possible with this analyzer. Well it is as it turns out..kinda sorta. I built up a simple test fixture to accommodate these bandpass filters:

And here’s a few plots demonstrating the results of the tests:

Both of the above plots have an approximate resolution of 200 KHz / div. The vertical has not been calibrated as yet.
This has been a fun project and interesting learning experience. I’ll revisited this project once it’s situated in the chosen enclosure…

As always thanks for visiting…Randy

Leave a comment

Intermission: Thoughts on an Upcoming Bench Project

Back in the mid ’80s I had an opportunity to become somewhat familiar with a Tek semiconductor curve tracer we had on a back bench in our lab. This was a Tektronix Model 576 originally introduced to the market in 1969 at a touch over $2000…not cheap

I thought at the time how nice and useful a similar instrument would be on my home bench. Of course the price tag alone precluded such a commercial version as this. But, I did find a few projects in the Electronics magazines published back in the 60’s and 70’s at the local library. All well and good but in reality I didn’t really have the time or resources then to realize a homebrew curve tracer of a type I would find usable.

Here’s a few links to some semiconductor curve tracer projects and info back when:

1. Radio-Electronics November 1965 “Component Curve Tracer

2. Radio-Electronics December 1969 “Transistor Curve Tracer

3. Electronics World August 1971 “Transistor and FET Curve Tracer

4. Radio-Electronics June 1972 “Build a Transistor Curve Tracer

5. Popular Electronics June 1972 “Build a Versatile Semiconductor Curve Tracer

6. Radio-Electronics November 1972 “Using a Solid-State Curve Tracer

7. Radio-Electronics April 1975 “Using the Curve Tracer

8. ETI January 1979 “Curve Tracer

A slighter more current curve tracer project that caught my eye is located here
Popular Electronics’s May 1999 “Semiconductor Tester” project…

This project seems to be a nice compromise between functionality and versatility. But first there’s more work to do on the spectrum analyzer…

Next time…Randy





Testing the MC3356 Spectrum Analyzer Part 7

First of all I’d like to know if anyone of my readers has attempted to duplicate this spectrum analyzer project, and if so, what your experiences have been?

OK so this past week I’ve finished up and integrated the 15 KHz filter module onto the RF main card and had a quick look observing the video out with an 500 KHz square wave input signal. Below are a couple of plots illustrating the quite substantial difference between the bandwidth settings:

500 KHz Input / 250 KHz Bandwidth

500 KHz / 15 KHz Bandwidth

The 15 KHz setting is quite narrow…here’s a few additional plots:

Antenna Input / 250 KHz Bandwidth

Antenna Input / 15 KHz Bandwidth

I think the 15 KHz filter is worth the extra time investiture depending on the flexibility required for this measurement instrument. I’ve now begun designing a front panel for the SA so I can start the sheet metal aspect of the project which I’m looking forward too…just need to fit it all in the box.

So that’s the story right now…it’s been a busy week. Saturday was spent traveling to visit my eldest Daughter and taking care of some home repairs etc.

Take care…Randy


Leave a comment

Testing the MC3356 Spectrum Analyzer Part 6

Wow! There’s been quite a few of posts in this series… A lot of my bench time seems to come in the very early hours of the morning. Working full time along with family responsibility’s can fill up a day very quickly. Lately I’ve been vacuum cleaner repair guy and with Summer in sight I’ll be putting on the lawn maintenance hat. And what else…Oh yes; I’m currently on my backup PC since my current one’s power supply seems to have died expectantly a few days ago…need to get a replacement soon. UPDATE – Done!

OK, so where are we with the SA. I finished up the 15 KHz filter module this week, but have yet to incorporate it on the RF board. Also, I began putting together an audio output / detector board that I’ll complete and test out sometime down the road. Here’s a few pic’s of these cards:

15 / 250 KHz Module


Detector/Audio output card

I thought at this point I’d take a overview of the Spectrum Analyzer’s output with the bench’s analog scope since this really was designed with this generation of instruments in mind. Again, very little amplitude or frequency calibration has been done at this point in project construction…

500 KHz Input


500 KHz Input w/ faster sweep rate


And a look at a long wire antenna Input:

And finally here’s a little video describing the Spectrum Analyzer’s operation with the above scope. Excuse the audio, it’s not the greatest…


Thanks and I’ll see you soon…Randy


1 Comment

Testing the MC3356 Spectrum Analyzer Part 5

I‘ve now reached a convenient break point in construction (well sort of, I’m still working on the filter module) and I thought I’d take a look at current circuit operation and performance. Below is the current state of the analyzer hardware (a few days ago):

Here are a few notes concerning this state of build:

1. The 1st Mixer LO frequency range is way off (should be ~ 145 – 235 MHz range). The schematic and parts BOM in RadCom’s 1989 article is really intended for a particular layout. PCB’s were produced at that time for all circuit modules i.e. rf board, sweep generator and PS/marker generator. JAB electronics in the UK have had these cards at various times although I’m not sure if they have them in stock currently. Anyway, working with VHF frequencies the layout matters a great deal and particular values need to be customized accordingly. That is certainly been the case here. Whether it’s deadbug, manhattan, a combination approach or laying out your own PCB’s tweeking is in order. I think it’s best to start with stated values to use as a baseline and then adjust accordingly.

Here’s an interesting plot with an input of an 100 KHz square wave: Both halves of the spectrum are displayed, being either side of zero hertz (DC). This center point shown below corresponds to the 1st IF frequency (145 MHz) so named the zero spur. The zero spur would be positioned at the sweep beginning and here clearly is not the case.

Below is an additional plot with the zero spur in approximately the correct sweep position:

2. I flipped the 2nd 10.7 MHz ceramic filter and the signal leading edge (near peak) distortion disappeared (good!). Certainly the overall signal shape is a reflection of the 2nd IF filtering. The overall noise level seems currently to be higher than an earlier iteration. I did replace the Video out cable (RG174 coax) connecting the RF board and Sweep board video amp with just a convenient easy hook jumper. Unsure if that makes a difference or not…One thing I noticed was my LED bench lamp added a bit of baseline noise too. It also helps to reduce the scope’s BW to 20 MHz. Here’s what we have now with an open ended BW setting:


3. Sockets! For VHF constructional practices sockets are a no-no. A better option here would have been to use a .1 center carrior card, bend the IC’s (MC3356 and NE602) leads at right angles, solder and eliminate the sockets for a tighter build in these areas. I did not on this build and even thou I strived to remove excessive lead lengths using the shorted paths I could, such as using SMT 1206 NPO caps where appropriate, reaching the higher frequency’s will be difficult. Something to keep in mind.

So what’s next:
1. Arrange RF board and sweep board as they will be placed in their final chassis positions. Cut lengths of RG174 when appropriate and prepare and install. Add any extra length needed for routing.

2. Currently I’m using an open coil for L2 with 3.5t which seems to get in the LO frequency ballpark. I would prefer better coil range adjustment however so I’ll wind one on an adjustable core form. UPDATE: Decided to continue with an air coil design and found that in my layout case 4t (22 awg) slightly expanded did the trick – Randy

3. Add a 2nd video filter choice (680pF) to see if desirable.

4. Finish the 15 KHz bandwidth filter assembly and integrate onto RF board.

Any questions or comments are always welcome!

’till next time…Randy