Almost exactly how I did it back in the days. I used a 8-12mm drill bit sharpened as a trident so that the central pin was a very bit longer than the side ones. It allowed the bit to center itself when operating. I used a simple hand operated drill to cut stuff. Later I switched to a piece of hacksaw blade sharply broken so that I could cut the rectangulars. I was 14-16 when I did the most of my experiments. Before the things went so messed up. And now I don't even have time for it which makes me feel really bad. I'm just working and working for my family to have a roof above their heads. After we had to storm out when that hell in Ukraine started. Thanks for the videos again - it's like a fresh air breath.
Thanks Alan! I am tired of Chinese companies locking MC's so that I can't save the data to program a new MC if for some reason the one in the device gets knocked out. So I built one of these Colpitts oscillator's to test crystals with! I have a whole bunch of 2N2222A's, so I used them. According to the datasheet, the 2222A has 50MHz higher transition frequency and 5pF less input capacitance than the 2222. It is working great!
Thanks for the excellent video example and explanation. Caused me to watch some of your other videos. Built the ckt and it works well with xtals in the 2.5-13mhz range. I wanted a 1mhz crystal calibrator and found I needed to change some values to get it to oscillate at 1mhz. Changed the 10k base to gnd to 6.8k. Changed 470pf B-E cap to .005uf. Changed 470pf Base to gnd to 820pf (a .001uf would probably work too). Works great. Now installed in a DX-160.
I used to painstakingly cut tracks in a circuit board with about a 3mm diameter diamond wheel bit but this looks way easier. Diamond holesaws can be purchased pretty cheap on eBay. Thanks for the video.
This was extremely helpful. I have been thinking about building some RF circuits and heard that Manhattan style construction was a good choice. So this was a good overview of the technique.
FYI, something I've found helps with the "grabbing" problem: take a slightly damp paper towel, put it on the drill press's table and then set the circuit board material on top of it. The moisture in the paper towel makes the entire apparatus much steadier which helps with drilling the islands; It also makes cleanup a snap.
This was great. I've been looking for a video that went thru the whole process of a Manhattan style circuit construction. Thanks so much for making it.
Thanks for the schematics. It worked well. Some of the crystals showed nice and other distorted sine waves. I managed to rip off the protective canister using an iron file, sharp wire cutters and a thin flat screw driver, as can openers. Aiming a 5mW red laser pointer beam (with laser class glasses protection) on the bare crystal turned the output signal unstable without changing the basic waveform. Also tested resonators, some of which produce square waves, and managed to open the can on one of the miniature oscillators in HC 49S packages to see how nice and delicate is the crystal.
Just saw this Alan; I wish I had it for tech students doing RF projects - excellent, exactly as it should be done. Another method - I got a metal punch to make and glue PCB islands to the board but it is not as easy or nice looking. Always a pleasure to watch. We use your videos now for our ham classes. One habit though that I encourage is to connect ground clips first - to prevent ESD and also surprises when measuring HV. Cheers
I have only just started having a go at building RF circuits using islands or pads. So far, I have found some pros and cons with each method. The pads have lots of advantages as they are easy and quick to glue onto the copper clad board. They can also be moved or removed if required. Additionally, they help to raise the components off the surface slightly (reduces risk of shorting to ground) and helps to keep lead lengths really short (reduces unwanted capacitance and inductance). You can also use very small pads that take up less space than islands. The disadvantages are that you have to make or buy the pads. Also, if you pull, twist or bend leads of components already soldered to a pad, it can become unstuck. So, obviously, big advantage of the islands is that they are permanent (though this may be a disadvantage during development) and, once you have a cutter you can drill as many as you like (at least until the cutter wears out). Use the method you are happy with or best suits the task I suppose. Just some thoughts on what I had found so far and certainly welcome any thoughts on pads vs islands.
This case caught my attention because i had a bunch of unmarked chrystals lying arround and i didn´t even knew if they would work or not. So i built the oscillator. I didn't use the island cutter method (used a common prototyping board) but i found the method and your expanations extremely good. Probably i´ll give it a go in the future. Anyway, in the end i found out that a lot of those chrystals i had didn´t work, even some marked ones... It was a lot of fun and useful stuff. All this to say: Thanks.
For those (like me) who don't have a drill press , take a piece of wood about 1/4" thick (and long enough to hold) and drill a hole in it about the size of your diamond hole cutter. Press the piece of wood on top of the copper clad board with the hole over the point you want to cut your island. You then have a guide through which you can pass the diamond cutter.
I tried this today and it worked great and it was fast! I glued little squares of copper-clad for islands but going to buy a rotary spot welder cutter bit from Harbor Freight tomorrow. Was able to check out several unlabeled crystals.
I like the island method. It is easy to set up. I already have diamond circle cutters of assorted sizes, which were made for cutting holes in tile. This is an improvement over the dead bug method because it is easier to visualize the schematic with the island connections. It's easy to add features to support testing and troubleshooting. No delays for pcb layout and prototype manufacturing... And it can even be used for RF work. Very, very nice! Thanks w2aew.
I did actually start to build a power supply that I have been postponing for the last 4 months probably. Although you could get a second-hand professional power supply on Ebay for around $40, building it is a lot of fun. I thank you for the inspiration sir!
I keep re-watching your tutorials; every time it is a new experience. Thank you, Alan. Can you please continue the series of the Basics? Happy New Year!
It would be interesting to make a qualitative analysis of the performance as an oscillator by observing the shape of the trace on the oscilloscope and a measurement of the output power. I've built one (different capacitor values, 680pF) and crystals up to 9MHz show a distorted sinusoid, like the one you get, more or less, whereas with 12, 16 and 20MHz I get a practically perfect sinus shape.
excellent video, i really enjoyed it! i've only utilized the toner transfer etching method, but i really like the simplicity/minimalist approach to this prototyping model. i want to try it asap.
Do you buy your boards with single-side or double-side cladding? I'm thinking about building a (relatively) simple analog synthesizer, and I want to use single-clad board with the circuit on one side and the front panel on the other. Thanks!
Typically - I pick up whatever I can find cheap! Usually, it is double side cladding, which is good for the RF circuits I make. You could use double sided for your project - just polish up the copper with steel wool, sandpaper, etc. and coat it with urethane and it will make a nice brushed copper metal finish to your front panel.
I built this circuit, and it works. I didn't have my yuppy 'Upper East Side Manhattan' hole saw handy. I just cut up a grid of small 5mm square islands from the copper clad and hot glued them 'Haarlem 'n Handy' to the ground plain board. Hope your feeling better Alan and the ankle is healing quickly ;)
Nice clean assembly & test. I noticed the waveform isn't quite symmetrical. Is that because the xtal is overdriving? I would have put a DC blocking cap between the socket and base. This allows testing parallel resonant tank circuits without shorting the base bias to ground.
These videos are outstanding. Excellent conversational teaching style. Question: I noticed that your oscillator waveform has some higher order products in it as you test the two crystals. Do you cover this topic in another one of your videos or was that a characteristic of the crystals you tested?
This is very typical of simple oscillators or this type - not so much due to the crystal, just the nature of the simple single-transistor oscillator circuit.
hi ***** great video, I had one quick question, The transistor needs to be fast enough for the crystal right? should I get a specific one to rebuild this or do you think most transistors will do just fine.
The transistors suggested are written on the paperwork near the osc design. In general you will need a transistor with a transition freq, Ft at least twice the desired osc frequency in order to be sure the osc. will start. In fact the true start conditions are a little more complex than that and they are tabulated in old TI semi books for Hartley and Colpitts designs especially. Audio transistors simply will not work in this circuit -they have too much input capacitance and too much capacitance from the base to the emitter lead. One transistor that simply shines for Osc duty is the BSX-20. You may be able to get it from Futurlec.com . In Oztralia , Rockby in Melb used to have stocks. It has a metal case so it keeps cooler when driven hard and the Ft is about three times better than the generic 2N 2222. Datasheet njsemi.com/datasheets/BSX20.pdf. If you have some marginal microprocessor grade xtal and the circuit will not osc with a 2N2222, then putting in a BSX-20 will often make the difference, esp between about 16 Mhz and thirty Mhz.
@@ernieschatz3783 Ft is the frequency where the transistor''s current gain is equal to zero. Obviously this statement is simplistic as current gain means something only when the topology , eg common emitter , is known . Therefore the Ft for a given part from a given mafr is defined by his measuring circuit topology and this will be detailed on his datasheet, one hopes. There is more detail on this in Wikipedia under gain-bandwidth product en.wikipedia.org/wiki/Gain%E2%80%93bandwidth_product. This article suffers from some crappy editing - omega C is not defined but it appears to be the - 3dB drop angular frequency for a given op amp (in which case the Gain BW product is defined in terms of voltage gain).
@@ernieschatz3783 you might like to modify this circuit slightly to work with a J310 nFET (which works at VHF). The bottom 10 K bias resistor on the first transistor increases to say 25K ohm. The upper one disappears. You also had in s small signal diode across the new bias resistor to GND to soft clip the RF going to the gate of the FET ; this then fixes the magnitude of the amplification. BTW the 470 p capacitors produce an impedance transformer effect from the 1K emitter resistor of TR1 back to the base circuit of TR1. Such a circuit will be more repeatable. You might find TR2 , the OP transistor has to change to a high gain transistor like an MPSH10, because the FET will need to see a higher input impedance as its load. (might find etc - these things are calculable or can be tested/trimmed up by simulation on LTSpice, which you can download from Linear technology ) There are only a few places to go for RF parts at a reasonable price nowadays but you can try Cecil the Parts place (do a google search on it) - he also sells magnetic components such as toroids and pignose baluns which are useful for RF transformers. There is or used to be a whole PDF of interesting AF and RF experiments by Todd , a Canuk ham. qrp-popcorn.blogspot.com/p/blog-page_9.html then go to the download link, not the NT7S Blog - of course quarantine anything you get and virus check it (I would but there was no virus). It is just one big PDF of all sorts of experiments and theory. Every circuit is illustrated with a point to point wired PCB so you can trust, to an extent, that he built the damn thing. You will find JFET bias, RF power amps .. all covered there one way or t'other.
Yes, you can do that. Often times, VCC are plane layers in a multi-layer board. In this case, you just have to take care not to short VCC to anything (chassis, etc.).
Hmm? So that is the Manhattan style? Okay then, interesting, and quite effective. Not quite as crude as "dead bug style". I was etching boards with ferrichloride when I was about 10 years old. Any lacquer based marker makes a resist pen for the etching fluid. The "Industrial" Sharpie is one that works really well, and for RF shielding dead copper / large ground planes, just brush on some lacquer paint. My first hand rendered through hole PCB artwork, taken from a schematic, for etching was the Fender fuzz wah/volume & tone pedal for guitar. It had an op amp, or two, and a couple transistors, & all the passive stuff. Not a great choice, but it works, more or less. The combination rocker & twist pedal case proved impossible to locate or build without some difficulty.. My very first PCB was taken off a photocopy, of the PCB traces & pads. It was a bootleg pay television descrambler. I used dry transfers on it. What a tedious pain in the neck, only lower! 73 KI7AQJ
Do I need a frequency counter to test a circuit like this, or would my digital oscilloscope do the job? (it does show frequency in measurement mode as well as pp voltage etc)
The scope will be perfectly fine. A counter will give you a more accurate frequency measurement than the scope, but the scope is certainly sufficient. Of course, you can also see the signal quality and amplitude on the scope - these are things that the counter won't do for you of course.
Thanks for the instructional oscillator video. What was more interesting was your “islands” that’s a clever means of making a circuit board. Special thanks for that. The other item that was interesting was your Soldering iron. I’ll appreciate, if you could let me know the make and model of the same. Cheers.
Howdy. Nice. No intention to be a wise guy but I think You made a good island chart on the schematic. I would have used that. Personally I like the layout to be like the schematic. For me it facilitates easier fault finding and helps keeping my thinking straight. Regards.
Another cool little useful circuit, you keep on giving these nice treats to us. Thank you very much for sharing. I probably going to put this circuit together myself on a protoboard because I have a bunch of crystals too that I've taken off some old stuff that people throw away and I have no idea if they are working or what frequency either so having this circuit is pretty handy to have for testing the crystals. If I keep going like this I'll have all these nifty circuits from you for myself and I think I'm saving a lot of money by making these circuits myself from junk parts, some parts are new but very cheap, however with these circuits always handy I can do a lot of different things already and I don't have to go and search and spend tons of money on professional equipment that do pretty much the same thing. Man, I feel lucky to have found your RU-vid channel. I wish you have a great day. Ricardo Penders.
This is probably a dumb question but lemme ask anyway: when designing a ccircuit how do you know when to use electrolytic or tantalun or some other kind of capacitor?
Not a dumb question... ...the choice of capacitor type is usually dictated by the application and value. If the application required a large capacitance value, then that will usually mean electrolytic or tantalum. If low ESR is needed, then a tantalum would be used. If cost is the driving factor, then an aluminum electrolytic might be chosen. If the value is small, then it could be a multi-layer ceramic cap. If low leakage is important, then it would probably be a film cap. It all depends on the application...
Built the oscillator on a prototype board, did not have the exact NPN transistors and did not have 470pf caps. I used 330pf caps. Tried a 5 MHz crystal and it worked but the sine wave is not pretty. I ordered some 470pf caps And the recommended transistors. I have a 27.185 MHz CB crystal. I want to build a test oscillator to check the receive of CB transceivers. When I used that crystal, the oscillator had issues. Should I assume the 470pf caps help in cleaning up the sine wave?
Will it also determine the resonant frequency of a piezoelectric diaphragm (buzzer). I tried, it didn't, even though I put larger capacitors because its resonant frequency must be around 10kHz. Maybe I am doing something wrong with the circuit, or is there a fundamental difference between quartz resonators and piezo buzzers? How to modify this circuit for measuring the resonant frequency of a piezo buzzer?
In this case, I did obtain a small pile of small PCBs (from QRPme as part of their Manhattan Chowder kit). But, I've had good success wil a few methods like: scoring both sides with a knife and breaking, using tin snips, using a a jig saw, etc.
It is very difficult to make one circuit that will work over a very wide range of frequencies. The values shown seem to work well for crystals in the lower HF frequency range. To operate at higher frequencies, you'll have to lower the two 470pF capacitor values. For much lower frequencies, you may have to increase these capacitor values.
Great tutorial. Thanks, as always. I would like to see an identical circuit built with various construction methods, then check the output of each for differences. i.e breadboard capacitance effects etc. I would like to do this myself, but my test equipment is not up to yours for spec. Thaks again Stoneslice.
Hi Alan, Great video! I am trying to create a circuit a nebulizer a scent diffuser but am not having much luck searching around. I'm wondering whether this circuit still works with a piezo plate in the place of where the crystal is? Also it is tough to find much info on piezo plates, is there a certain range of current going through them that they work better? Thanks!
Im not an expert on ultrasonic piezo drivers, but a quick google search reveals lots of good results: www.best-microcontroller-projects.com/ultrasonic-atomizer-circuit.html product.tdk.com/en/techjournal/tfl/sensor_actuator/NB/ www.murata.com/~/media/webrenewal/support/library/catalog/products/sound/p15e.ashx www.senscomp.com/pdfs/using-piezo-sensors.pdf www.techbriefs.com/component/content/article/2-ntb/tech-briefs/semiconductors-and-ics/5615
Great video as usual... Sorry if this is a basic question but... What would be the easy/simple way to "clean-up" that output wave ? I guess it's not about applying a LPF because it should work on any freq... maybe you need to make another nice-sine oscilator and "drive it" with the first dirty oscilator in order to get only the fundamental freq ? (I'm not an EE... just an Amateur trying to take a guess here)
A low pass filter or tuned circuit load that only passes the fundamental would clean up the output. Of course, the design of the filter/load will be dependent on the crystal frequency being used.
Hi Alan, I tried to run on 72.310MHz quartz, and it's executed at a frequency of 12.5. As known quartz can only operate at odd harmonics. Why could it happen?
Do you have a video or know where I can find info on how engineers know where to place what components when designing? I have formal training in troubleshooting circuits but it’s always baffled me how engineers know where to put everything.
What range of crystals does this work for? I built this and it seems to only work up to about 12Mhz.. Any of the 20M+ crystals I have don't show up... But I see you using a 60M if I recall..
I use an island cutter with a center pin I made from a flathead bit (just filed two slots into it and sharpened the edges.) The pin goes into holes I pre-drill. It's self-centering and one doesn't need a drill press
I've got an old Lafayette model HA800 short wave radio that is dual conversion with 455 Kc and 2.608 Mhz IF's that quit working after some parts in it's 2.153 Mhz second osc. drifted off freq. It's a Colpitts osc. like the one you used in the video so I built it up on a bread board to get it working but I found that some of the parts values they originally used wouldn't work so I put some new parts in and mounted it back on the radio's PC board. The problem I'm having now is that since the output is not a sine wave but more like a saw tooth similar to the one you showed, I get tons of harmonics all across the dial! Do you know if there is a way to get a better sine wave out of this circuit? Thanks.
The most popular ways to get lower harmonic distortion with simple oscillators like this are: - Use a low pass filter on the output - Use a tuned-circuit load in the collector - Tweak the gain of the circuit to be just enough to sustain oscillation
@@w2aew Thanks, I'll try that. Oddly the radio did have a tuned circuit across the oscillator but the parts values are way too small and it is resonant at 15 Mhz instead of 2.153 so it needs a larger coil to make it work. The radio seems more like a science fair experiment than a commercial product but I guess that's why they came out with a version A then a B!
Ordered the cutter. Tried three different pcb materials. The cutter removes the entire dot. I’m also seeing that the removed copper remains in the cutter. What’s the trick?
Awesome work. It would be really cool to have an island cutter that also drilled a through hole in the center. That way you could just stick the leads through and solder them up trimming afterward. With a double sided board you could even use a combination of drilled and not drilled to make the back side a power plane..... Maybe
I'm picturing three bits. One that makes just the island, on that makes island with through hole for connecting multiple things together also to power and one that strips the island away with through hole for single power with no other connections. A double Manhattan! (I have no Idea what I'm talking about btw:)
You could probably do this with a traditional brad-point bit. They typically have a pointed lead-in and cut the perimeter shortly after that. Even if the lead-in doesn't go through, you would have a starter hole to run a second small bit through for the lead hole.
When I was young I used to use tag-strips salvaged from old valve televisions. Compared to Brattian & Bardeen's first transistor I reckon my efforts were pretty tidy. :-)
Can you tell if this work for 6 or 10 different christals or we need for each christal another plate what is calculation sheme for ten christal do you know that
hi..its a nice video ..what is peak to peak voltage of op waveform and what if i use 40mhz crystall..will peak to peak change ?? and also can u suggest method to amplify op
What frequency range of crystal oscillators will this test?Nice job on the build.I need to find my pad cutter tool I used years ago.Guess If I have to I can cut pads and glue to the copper board.
Thanks a lot for the video, i was inspired to create this circuit. It functions nicely, but i have one query. The output is not a pure sinewave, somewhat distorted. What step can i take to make the output smooth ?
Simple oscillators like this have a lot of harmonic distortion. The best thing to do is to use a low-pass filter on the output or use a tuned amplifier on the output.
@@w2aew I was also thinking something similar to use. You have now confirmed it. Thanks again for the reply. I really appreciate your efforts to support DYI community.
@@minazulkhan8287 He shows you the scope that is probing the output. It shows the signal output is about 5V p-p. The transistor will lose gain as the frequency increases, so the level will be less at higher frequencies. You'll need to make sure your transistor will work at 40 MHz, and you may need to use lower value capacitors.
I don't understand...it looks like you are cutting circles in the copper thus isolating it from the larger area which is ground. I get that but it sure looks like the solder is breaking out of the circles or islands and touching the ground plane. How do you avoid that?
Due to surface tension of the molten solder, it does not cross over out from the circle. just leave the tip of the soldering iron on the circle and do not let the tip bridge the gap. It is easy with a little bit of practice.
Is it safe to add untinned component leads to lands that have already been soldered without adding any extra flux or solder? Don't you get problems with the old solder not wetting to the new lead?
Most of the component leads are already tinned, so it isn't a problem. If the leads aren't tinned, then I'll usually pre-tin them before soldering. If I have a problem with wetting, then I'll apply flux.
No. The capacitor values are clearly marked in the schematic. Very readable at the 1:00 point. For frequencies above 10-15MHz, you may want to lower the value of the two 470pF caps to 100 or 150pF each.
Thanks for the quick reply. Higher freqs are definitely tricker to tame. Harmonics do become an issue. Thanks for the video its a fantastic practical guide for some experimentation. :)
what is the use of the upper 10 nf cap below the 100 ohm resistor? decoupling or ac ground to prevent signal loss in vcc? the 100 ohm to make the path more resistive to high frequency? or just for the biasing scheme?
The 100ohm and 10nF cap provide AC isolation from the main power source. It keeps any noise from the power supply from getting into the circuit, and prevents the circuit from injecting too much noise on the power supply.
+Horacio Ricardo Drut Sure it will. Mine works for most crystals down to less than 3V supply. At lower voltages, you may need to tweek some component values to increase the bias of the transistor.
+w2aew Hi again, I'm trying to make it work @ 3.3V with no success... Have lowered the biasing Rs to 3k, bypassed the 100ohm resistance and increased output C value to 1uF... What are the values you used? Thks for your help!
+Horacio Ricardo Drut Mine worked, as shown in the video, down to 3V or less - no changes in component values. Oscillators tend to be fairly layout dependent - so careful layout (no breadboard) is important.
+w2aew Hi, I thought I'd let you know how I improved the performance of this circuit. At least for me, I was able to increase the output p-p value by adding a bypass cap across the 1k resistor that connects both stages. Hope this helps someone else. Rgds.
