DIY rebuild of the Tek SG505 V3

This Blog post will detail the third stage of my project to rebuild the Tektronix SG505 instrument.

Here is the link to the post describing the second version of the design: 

https://www.paulvdiyblogs.net/2025/08/diy-rebuild-of-tek-sg505-instrument.html

Here is the first post about this project:

https://www.paulvdiyblogs.net/2025/03/diy-build-of-tek-sg505.html

The reason for the third revision

After the investigations of the second version, I wanted to separate the Power Supply from the Generator in the same enclosure. I also wanted to put the Generator circuits into a full metal enclosure to make the output of the generator as clean as possible and as a minimum, remove the mains hum.

After a very long search, this is the only enclosure I could find that has the required height and width for the front panel layout. It is a ProMa 130 0044 and is also available from Amazon. The outside dimensions are 165x110x80. I would have liked a black enclosure, but alas I couldn't find one. There is one available from ProMa though, with part number 130 0045. I can always spray paint it black myself if I develop the urge. The current Front Panel design will need some modifications and will replace the aluminum panel. The generator PCB will slide in a slot close to the bottom. Unfortunately, the dimensions are a little different so I can't use the current PCB in this enclosure, not even to try it out. 

This all means a new PCB for the Generator, for the Power Supply and for the Front Panel.

Splitting the Power Supply

There will be a separation of the noisy mains related parts of the circuit, that need to go outside of the enclosure for the generator. In essence, it means that the transformer, the bridge rectifier, the main capacitor reservoirs and the 40V regulator need to be on a separate PCB that will be housed outside of the generator.

This part of the project will be described in a dedicated post here:
https://www.paulvdiyblogs.net/2025/11/the-diy-sg505-mains-power-supply.html

The shunt supplies for the +/-16V rails can move to the main generator, and also the 12V supply can move to the main board. They are quiet and will have no negative effect on the generator. I hope. It also makes the interface from the Power Supply enclosure simple, because I will only need to use two wires for the 40V supply that feeds the other three rails.

The circuit after the transformer and the bridge will get some more filtering to avoid mains related noise getting into the generator.

The Power Rails on the Generator PCB

These three rails are very quiet by themselves and can now move to the generator PCB.

No major changes from the previous design, I just added a few extra capacitors and ferrite beads to the power input lines. This may still change a little based on the new layout.

Just when I was about to order the PCB from PCBWay, I decided to skip the idea of using a heat sink for the LM317, because it will still get too hot and raise the temperature in the box. It will now be mounted isolated to the side of the aluminum enclosure.

The Generator circuits

The other circuits stay the same, will just get a revised layout and incorporate the three power rails and needs to fit in the new enclosure.

I finished the new version of the generator PCB, now with the power rails on it.

This is what is ordered. On the bottom part you can see that the LM317 is now flipped around and moved closer to the edge of the board. It will be mounted isolated on the side of the enclosure to remove a considerable heat source.

The Front Panel

This is the updated front panel fitting the new enclosure.

The golden rings around the holes connect the front ground fill to the back ground fill to add an EMI shield to the inside circuits. 

The rings on the back are larger so will connect to the metal parts of the switches, connecting them to the shield. The 4 mounting holes in the corners also have exposed holes on the back and will connect the shield to the aluminum enclosure. 

The enclosure itself is not connected to earth ground but floating. I have created the possibility to connect the GND of the PCB circuits close to the output BNC to an exposed pad on the front panel. In that case, the circuit GND will be connected to earth GND when the BNC is connected to a DSO.

Building up the boards

I received the shipment with the three PCB's and I have built up the power supply by transferring most of the parts from the old board, added the additional parts and tested it. No problems.

The next step was to add the solder paste droplets to the main generator board and transfer the parts one by one from the old board. I used my heat gun to remove them and put them on the new board. When that was done, I reflow soldered the board. Because I used smaller solder paste droplets this time, the reflow process went a lot better, with only a few tiny solder ball bearings and a lot less of the flux gue. I did not clean the board just yet, I wanted to test the functionality first.

Discovering issues

Bad Solder joint

At first I wanted to check the voltage levels of the three power rails. The +16 was only about 9V and then dropped to 3V, the negative 16V was about 30V, the 12V was OK. Although nothing got warm, I quickly shut it off. After connecting the switches and potmeters to the connectors such that the generator could function, I applied power again and saw a welcoming sinewave, albeit with some distortion at the top half. Hmmm, partial good news. When I checked the 16V rails, I still noticed a large unbalance, and that explained the distortion. The good news is that the most complicated circuit seemed to work OK, but the most simple circuit did not, but why?

The hunt for the shunt supply imbalance turned out to become more and more strange. To a point where I started to remove the parts from the shunt supplies one by one, but without any improvement. Using my Lab Supply instead to first power the +16.5V and the -16.5V everything worked, I then supplied the 40V supply, further up-stream and that also showed the correct currents and the system worked fine. Now really puzzled, I used another one of the spare generator boards and started to add the minimum amount of the same parts I just took from the board for the shunt regulator to make it function, which it did flawlessly. Even powering the generator from the second board showed the correct balanced voltages. I was flabbergasted. After thoroughly cleaning the board and resoldering the components back to the board in pairs, everything worked. Bad solder joint!

Wimpy transformer

I also discovered that the mains power supply had a wimpy transformer, so that also needed attention. Details are in the other post: 
https://www.paulvdiyblogs.net/2025/11/the-diy-sg505-mains-power-supply.html

Mistake on the front panel

When putting everything together, I first mounted the construction for the main potmeter and the reduction unit. It fitted perfectly, unlike with the first front panel. So, happy with that result, I added all the other switches and potmeters and proceeded to slide the board into position, when I hit a barrier.

Turns out that I made a serious measurement mistake with the position of the rotary switch for the multiplier. It was bumping to the board and also bumping to the reduction unit. Moving the hole up and left solved that issue, but it will mean another turn of the front panel.

This is how it looks now, so close...

This is the inside view of the now fully working instrument:

Mains related hum

When I did the first FFT tests, I still saw some 50Hz hum and some harmonics. When touching the metal parts of the front panel, it got sometimes worse, sometimes better. I did not have the main potmeter knob mounted, and when I touched the metal axel, the hum got worse. Connecting the Earth GND from the output BNC to the metal parts of the main potmeter and reduction unit with a wire did not do anything, but connecting it to the common GND of the main board reduced the hum dramatically.

It turned out that the mounting holes for the main potmeter support and the reduction unit did not connect the metal parts to the common ground of the PCB. I used star washers on each support to improve that. I also added a blank ring to the layout around one more hole of the contraption to improve that going forward.

Connecting earth GND to common GND?

I intentionally connected the front panel shielding to the metal enclosure to create a Faraday cage, but I separated it from the common GND of the board.  

In my current setup, with the USB connected EMU0202, however, that produced too much unwanted mains related hum. 

I already added a solder tab on the back of the front panel as an option to make the connection possible. When I soldered a wire to it, and connected the other end to a solder lug I added to one of the supports for the main potmeter to connect the two GND's together, it solved the hum issue completely. But now the instrument is earth grounded through the EMU to the laptop, which by the USB-C cable to the power supply is connected to earth GND.

The other possible connection for the instrument to get earth GND connected is through a BNC cable to a CRT or DSO and that will connect it to earth GND as well.

The original SG505 has a switch to connect the common GND to earth GND. If you also want to have the option to separate or connect the enclosure from earth GND, you could add a toggle switch to the back of the unit, or a sliding switch on the side. I'm undecided at this moment, but it's easy to add afterwards.

Result after the fixes

After all these mishaps and corrections, I wanted to share the first FFT from the generator, hot of the press. Note that I was able to quickly trim the second Harmonic visually into oblivion (0.00002%).

Result, no hum, no noise. 

Unfortunately, with H2 visually gone, H3 is now sticking out, but the rest of the harmonics are virtually invisible. 

During my testing with the updated power supply and mounting everything on the front panel a few times while fixing things, I noticed that I could no longer adjust the H2 harmonic as low as it was above. There is now also some hum visible, so when the new supply arrives, I will look at it again in more detail. 

Just for reference, my DIY version of the generator seems a bit better than the original one. Albeit using different measurement tools.

I'm almost there...

Building the final version

Happy with all these results, I uploaded the updated power supply V4 and the updated front panel to PCBWay for production. They gracefully continue to sponsor my activities, despite my screw-ups.😇

It will usually take about a week for the shipment to arrive.

I got the front panel, but made a mistake ordering the power supply board, so that will come in another week. See the dedicated Blog post for that part of the project here. 

In the meantime, I started working on the various BOM's, including an off-board one for the front panel parts. They, and all other information will be on my Github site that I will publish when I'm done. 

I have been using a lot of parts I already had, but needed to find parts that I could put on the BOM so others can order them. Here are some details for the off-board parts.

Output signal potmeter

One of the challenges was to find a potmeter for the output volume with a switch, that is activated at the end of the rotation. Most of the ones I found switch at the beginning, at 50 degrees according to the specifications.

I put one together with parts I had from a previous life, and that looks like this:

These potmeters come in separate segments you can take apart so you can create stereo versions, or use two different resistor values, and optionally also add a switch segment. I was able to turn the switch segment around to get the right action. To have the switching indent at the end of the range is more natural, because the switch fixes the generator output by a resistor divider just below the maximum output. This is also how the original SG505 has it. I put a potmeter in the BOM with a SPDT switch, but it will be activated in the beginning of the rotation. If you are able to find a potmeter with the switch activated at the end of the rotation, let me know in a PM so I can share it with others.

Mounting the main board

You can add all the components on the main board, but do not solder the 12V regulator in position yet. The height of the regulator to the board needs to be determined first. 

It's now time to mount the main potmeter and the reduction unit using metal brackets to fix the position of the axel in the middle of the hole of the front panel. Add a solder lug on the screw that fixes the reduction unit with the screw hole that has an exposed ring on the bottom of the PCB. The is the one closest to output on/off switch connector. Lightly tighten the screws and nuts and try the position of the whole contraption with the hole in the front panel. When you're happy, see if you can tighten the nuts while the PCB is still on position. If not, you need to carefully slide the PCB a few cm out of the enclosure slots such that you have access to the screws on the bottom. Fix everything and try it again.

When that is done, you should have the main board flush to the front panel, with the potmeter axel in the middle of the hole. 

To mount and fix the 12V regulator, we need to drill a 3.5mm hole in the side of the enclosure to give it a good heat sink. Position a sticker on the inside of the enclosure about in the position where the regulator will be. My hole was drilled in the middle of the third rib from the bottom. The hole should be drilled the middle of that cooling rib, because the head of the screw need to be in the center between the ribs. Position the main PCB in position, such that it will be flush with the front panel. Put the LM317 in position after you have clipped a few mm from the legs so they don't connect to the bottom. With the regulator inserted in the PCB, note the position of the hole and mark it on the sticker. Use a caliper to measure the distance from the front edge of the enclosure to where you marked the center of the hole of the LM317. Transfer that measurement to the outside and mark the position in the middle of the rib. Remove the main board and drill a 3.5mm hole in the enclosure. Grate the hole on the inside so there is no burr. Slide the main board in position again, and maneuver the hole of the LM317 to align with the hole in the enclosure. use a screw and nut to position the LM317 and solder one or all of the pins to the regulator is in the correct position. Remove the screw and carefully slide the main board out of the enclosure. You can now permanently solder all three pins to the main board. 

Important! When you will fully mount the regulator later, (not now) you have to use an isolation pad (mica or silicon) and a plastic isolation ring to make sure the metal pad of the LM317 is fully isolated from the enclosure, or you will have a dead short of the supply rails.

Fixing the OLED display to the front panel

Do not mount any of the other parts to the front panel yet. It allows you to handle it and the fragile display easier. 

I first used a black permanent marker to color the inside of the rectangle in the panel pitch black. Do that from the inside so you don't smear ink on the face side. 

Make sure you can power the display from the main board while positioning it so you can see where the text is, relative to the viewing area. It will be virtually impossible to position it correctly otherwise. If your main generator board is not inside the enclosure with the 12V regulator cooled by it, you have to use an external heat sink, because the regulator will get too hot otherwise.

Using glue

You can try to glue the OLED display in position to the new front panel. You need to add some glue (not instant glue) sparingly(!) so it does not flow into the visible area when you gently(!) press the two together. Add the glue on the front panel backside. Power the display so you can see what needs to be visible and position it horizontally, put it in position, and keep it there until the glue is dry enough.  I first tried that, but abandoned it.

What not to do

In my earlier attempt, I used a washing cloth clamp to secure the OLED display in position when letting the glue harden. We'll that seemed to have destroyed the display because it turned black. I first thought that the clamp must have pressed to hard and damaged the flexible cable connection between the glass and the PCB. I had no spares so needed to order a new display. 

Well, after the new display's arrived, none of them worked. Tongue in cheek and red faced I have to admit that it wasn't the display. While trouble shooting, I measured the SDA and CLK signals with my DSO and I measured 5V on the display PCB with my DMM, so I started on a rat race trying to find the issue. After a while searching, it turned out that the crimping of the GND wire on one side was bad and must have disconnected while clamping. Although visually it looked OK. Long story short, what I did wrong was that I connected the GND of my DMM to a GND pin on the main PCB, in effect bypassing the bad GND wire connection when measuring the 5V. I did the same with my DSO, but that still showed the presence of signals so I did not look further. I replaced the GND wire in the extension cable and did a better job crimping the connecter and all is well again.

In hindsight, it's probably safer to use some normal Scotch tape to keep the OLED display in position while the glue is drying. 

Using Scotch tape

I ended up using Scotch tape to secure the display, I used electrical tape earlier but that's too flexible and the display sagged down a bit. 

In the picture below, I used another PCB on top of the front panel so I could position the OLED board better. In my case, the distance between the top of the OLED board and the top of the front panel was 20.78mm. You can easily check the horizontal adjustment that way. Because I used clamps to press the two boards together, I could rather easily adjust the vertical and horizontal position of the OLED display within the rectangle opening. To do that you have to power the OLED display from the main board so you can see where the text is.

After positioning the display with one piece of tape while positioning it, I used some more Scotch tape to finally secure it in position. Do not cover the onboard regulator with tape. I used a piece of double sided foam tape to bridge the distance to the front panel en the OLED display where the connector is. It should help with the stress put on the rather flimsy OLED board sandwich with the glass display when you plug the connector in.

It does not look very pretty, but does the job and is not permanent.

Assembling the front panel

Do not fasten the front panel to the enclosure just yet. First mount the switches with their cable harnesses already in place, with the solder ends shrink-wrapped for sturdiness. Ask me why. Mount the Vernier potmeter and the volume potmeter. Solder an about 8-10cm thick ground wire to the solder pad on the back of the front panel, with the wire going upwards. 

Mount the rotary switch into position with the connected wires on the top side, away from the main board.

Get the front panel roughly in position to the enclosure with the main board already in place. Solder the other side of the ground wire to a solder lug you should have installed on one of the mounting screws of the potmeter delay unit.  Take one of the screws that have a ground ring on the bottom so there is a solid ground connection. Check with an Ohm meter for only a few Ohms max. between the front panel exposed holes to one of the GND test pins on the main board and the outer ring of the BNC connector.

You can now connect the wire harnesses from the 0/-10dB switch to their locations. The wiggling room is very tight which is why it should be clear now that you should not have fully mounted the front panel yet.  Then connect the output on/off switch connector. At this point, make sure the switch operation is correct, because you can still turn them 180 degrees to fix that. The other connections are not critical, so you can now mount the front panel into position with the two screws going into the bottom half of the enclosure.

You can now add the ring and nut for the BNC connector and tighten it lightly such that the main board is flush against the front panel. You should have tried this position already when you mounted the voltage regulator to the side of the enclosure such that you don't put too much force on the leads of the LM317. With the front panel mounted, you can now also mount the LM317 into position. Remember not to forget the isolation! Double check again that the metal tab is indeed isolated from the enclosure.

You can now connect all the other wire harness connectors and add the knobs into position.

Calibration procedure

Now it's time to add power and check all controls and that the switches are in the correct working position.

By now, you should have a sinewave output and the counter should show the frequency. If all the controls work correctly, it's time to let the unit warm-up for at least 15 minutes. After that, it's time to finely adjust the 44V main rail, while measuring the voltage on the main board using the +Vreg and -Vreg test points. Finely adjust the rail voltage with the trimmer on the main power supply board.

With that done, you can now verify the voltages of the +17V and -17V rails. They should be within 0.5V of each other and 17V. If not you can change the resistors that are setting the voltage for the TL431's. If you installed the optional trimmers, you can set both rails as close as you can to 17V and equal to each other.

To adjust the 2nd harmonic adjustment, you need to use an FFT to see the effect. With the unit now sufficiently warmed-up and the rails adjusted, you can trim the 2nd H trimmer for a minimum height or dB value. You probably have to use averaging of the measurement to see it well.

Calibrate the 0dB output level. Select the Cal switch setting on the output potmeter. Connect the output of the generator through a 600 Ohm in-line terminator to a DMM in the AC mode. Select the 1KHz frequency, because the DMM will most likely be precise at this frequency. Adjust RV2, the 0dBm adj. trimmer for a reading of 0.77459Vrms.

There is a bit more to it. The adjustment range of RV2 goes from just below 0dB to about +4dB. If you desire another output level, you can change the value of R47. 

Verify that the -10dB output setting results in about 0.4887Vrms or -4dB. This was intentional for me, because I already have a 600 Ohm in-line -10dB attenuator. If you want a real -10dB attenuation, you need to change the values of R50 through R55.

That concludes the calibration of the unit.

In the meantime, the new power supply boards arrived and I built on up. The details can be found here.

I now have a fully functioning setup and will show some more measurements soon.

Stay tuned for more information while I go about making some final measurements...

Here is the first one after calibrating H2 to 0.000034%. Total THD is 0.00033% not bad at all, and virtually no mains hum (-144dB), and that is with the cover off.

A Github repository is available here Note that it will be updated with information during the project when I have verified the correct operation. I'm still working on it, so there is very little information there at the moment so please be patient.