Hi everyone. Very, very long post ahead with pictures for your enjoyment and possible consideration.
Most people on this and other forums know me for two things: the various Yamaha mods that I have produced over the years and the rather politically incorrect language that I use to voice my opinions on everything. Well, this time we will add something else to the mix: my first and only Casio modification.
It all started with the itch that I got some time ago to have a keyboard with "real" piano keys. I searched Yamaha first because almost every piece of equipment that I have in my setup is made by Yamaha. I checked various options but I came up empty mainly because of the prices I found out there. I am cheap, I want to have the best I can for the least amount of money. I did find a piece of equipment that I liked a lot, the Yamaha stage piano CP300. It has a dead serious industrial look to it and an excellent feeling in the fingers. It also has a street price of $2000 and despite the fact that I found a brand new one for $1400 it was still too much for my rather half-empty wallet. Don't you hate it when those nice opportunities present themselves when there is a money shortage?
Anyway, while I was at the store checking the CP300 and other keyboards, my eye fell on a rather simple but elegant keyboard, the Casio CDP100 electric piano. I put my fingers on it and immediately I thought "that's a real piano" because the keys were heavy as they are in a real piano. I saw the "used item" price at $200 and I liked it even more. Then I went searching for detailed specifications and learnt that it features 88 piano weighted keys with 3 types of touch response, MIDI IN and OUT, a damper pedal input, 5 voices with layering, 32-note polyphony, some basic reverb and chorus effects with combinations, and 8W output per speaker. More characteristics can be found on the manual, available online by Casio. There is no audio input and no line output. Only a headphone jack is available and that is used as a "line" output if needed.
No bad but I would have liked the audio in/out capabilities. I kept searching online and downloaded the Casio Privia PX300 manual. Externally the CDP100 looks a lot (really a lot) like the PX300. That is always a good indication for mod nuts like myself: many times a company keeps the manufacturing cost down by making one mold for many models and then adding options and labeling as the price and model number go up. In doing so the electronics inside can be very similar among various models and that's where the game begins. I went back to the store got the slim and long "baby" and then took it home for testing.
The piano sounds are not bad at all and when combined with the strings you can get some very decent performance out of it. The main reason for the purchase though was the weighted keys. There is no "spring" feeling and as we will see later on that's 100% true. There are small hammers inside that give you that nice piano feeling and contribute to the unit's weight of 25 lbs. The CDP is made of plastic and if you play the keys hard you will hear them knock but you cannot have everything for $200, right?
The modifications/upgrades I performed were extensive and not all of them were satisfying in the end. They included:
1) Replacing the damaged speaker cloth
2) Adding dual tweeters and a crossover on each side for better sound
3) Adding audio inputs
4) Adding audio outputs
The first item of the list will not be covered here because it is not important. Basically I used some grey-colored cloth that I had (that was of lower quality than the original used by Casio) to replace the damaged cloth on the keyboard. It is a process that takes a lot of time and the result can be bad if you are not careful. There is information available online on how to replace the cloth of speaker cabinets. This one was a bit tricky due to the curves of the speaker grilles. I would grade my effort and results with only an 8.5/10 but I am OK with this.
The addition of the tweeters was not necessary but I could not resist. There was so much empty space in there and I had to use it for something. The audio-input addition was relatively easy but the audio-output was not. So here we go...
Well, before you reach the point to add the tweeters you must first remove some covers and side panels. The small bottom cover comes out after 17 screws have been removed first. I do suggest that you put the screws in small labeled plastic bags so that you know where they go. There are various sizes of screws used throughout the device and you do not want to mix them. Also there are some holes that do not have screws but if you take a few days to put the cover back you can forget which ones they are. I used small yellow stickers to mark the unused holes. So, after you take the cover off you get to see where that piano feeling comes from:
The picture shows small weights attached to the metal frames that extend from each key. They are balanced and as you move the keyboard during the upgrades the keys move around their mounting point. Even if you use the flat surface of a carpet as your workplace it would be better to have some paper or plastic foil on it for the keys to lie flat. Also keep in mind that there is a lot of white grease on the keys, most likely silicone.
Then you remove more screws from the rear side of the main bottom cover and a few more from the area below the key frames. Use a magnetized long screwdriver because the screws are deep inside the unit. Store them in the bags I mentioned above. Flip the unit over so that you face the top side and remove a few small and big screws from the sides in order to take the covers off. Depending on the mod you would like to perform it may not be necessary to remove all the covers.
The next picture shows the left speaker of the keyboard.
You have to remove the grille first and decide what tools you will need to open the big holes for the tweeters and the small holes for the mounting screws. Do you need one or two? I do not know. I used the Dayton ND20FB-4 with an RMS handling of 15W and a DC resistance of 3 ohms. Two of them in series total 6 ohms, which is too close to the 5.8 ohm resistance of the original speaker that becomes the woofer. They were on sale from Parts Express for less than $7 each. They must be connected properly so that they are not out of phase, so the "red" terminal of the first tweeter connects to the "non-red" terminal of the next tweeter. Here is how the same side of the cover looks like after the "operation":
However, as you may know you cannot just connect them in parallel to the woofer. You need a crossover. You could buy one or you could make one. I made my own with some inductors and some non-polarized capacitors. The inside of the cover looks like this:
For the crossover I chose to go with something simple: a Butterworth 2-way, 2nd order, 12db/octave network made with capacitors and inductors. I could have gone with something better but the goal was to get decent performance at low cost. Of course some engineering was involved with simulations and tests. For that I use a nice free piece of software by Texas Instruments: TINA-TI. It's easy to master the basics in less than an hour without reading the manual. Instead of the usual 70% cutoff point I went with 80% so that I could have a more "flat" response. The next picture shows the crossover and the simulated results using the exact values of the components as they were measured with an LC bridge.
The inductors and capacitors came from All Electronics located here in California. Each inductor is rated around 360 uH; I used a combination of inductors to achieve a value of 247 uH for the woofer network. It's a long story to explain why I chose those values; an even longer set of equations is involved. I know that it works because the end result is pleasing to the ears. The LC network for the woofer consists of a 247 uH inductor and a 3.5 uF capacitor, while that for the tweeter consists of a 4.81 uF capacitor and a 360 uH inductor. As mentioned above you need two tweeters in series to get a 6 ohm value so that it matches well the 5.8 ohm woofer. You cannot see much in the following picture but the two crossover boards are visible near the speakers and mounted on the top cover.
Now, why would I need audio in? Because I wanted to have a simple setup between the Casio and the Yamaha QY70, that tiny yet powerful sequencer. With a simple MIDI setup I could use the instruments of the QY and have them sound through the speakers of the Casio. I wanted to mix the Casio sound with the QY sound somehow and I wanted it to be cheap. So I took a look at the sound board of the Casio that futures small (but not SMD) components and I noticed some empty spaces and holes.
I decided to give a call to Casio and ask for the service manual. "Are you an authorized dealer, sir?" the voice on the other side of the wire asked. "Nope". "Then we cannot sell you any manuals or parts" was the response. Really? When I call Yamaha they ask what credit card I want to use, not if I am an authorized dealer. They even send me service manuals for free as pdfs if I get to talk to the right person. So, screw you Casio. You won't see me buying another instrument of yours for the rest of my life. Let me make it clear to you: the hardware is mine and I do whatever the hell I want with it. Get it?
Most likely they cared less and so did I. The sound board went under the magnifying glass in order to reveal its secrets to me. The first item was the amplifier chip: LA4636 by Sanyo. You can find the pdf of it online if you search. Here is a picture showing pins 2, 4, 8 and 12 modified to reflect the connections of the CDP100. The rest of the pins show the components proposed by Sanyo, not those used by Casio. It's not important but knowing which is left and which is right is very important.
Then I took a look at a small board where the damper pedal jack is located. I also took a look at the Privia PX300 user manual. Does it ring a bell?
You bet it rang a bell but along with it came an alarm siren. You will notice small squares next to the jacks. There is an "L" letter over there and that implies inductors or transformers. But what inductors? Well nobody knows exactly. If you take a look at the various boards inside the CDP you will eventually notice some small 2-winding transformers used in almost every input/output connection, including the MIDI signals. Each side of those transformers was measured with the LC bridge to be equal to 50 uH. Of course Casio would not sell them to me nor would they tell me what were the values of the capacitors that surrounded them. So I had to go to the extreme and make my own transformers.
Casio used toroid cores and because there were very few turns on each of the two coils, the high value of "L" meant that the toroid core was made of high magnetic permeability material. I had no clue what that material was. I checked around the workshop and found a bag with toroid inductors from Radio Shack, item 273-0109. You may not find them in every store anymore, it's an end-of-life item. Another item I used from the workshop was some magnet copper wire, 22 AWG, in two colors, so that it becomes clear which side connects where.
The picture above shows the process of making the transformers. The original magnet wire must be reomoved. Then each of the two new windings needs 2-feet of 22 AWG wire. In the end you have around 30 uH of inductance per side. Not exactly the 50 uH Casio used but nobody knows if the same transformers were used everywhere. Each transformer took 30 minutes to make by hand. Then I simply soldered them on the board along with some jacks I bought from ebay (read China); three 1/4" stereo pcb/panel mount headphone jacks and one 1/4" mono pcb/panel mount headphone jack (for the L/mono line out space). The three stereo jacks required one of their pins to be cut to fit on the board. I chose 10 nF capacitors to populate the board and that may or may not have been a good guess for the output part of it.
Would it work without the capacitors and the transformers? Yes it would. I do not know why they chose to use those transformers but I have a suspicion. Casio used a single power supply design and that can increase the amount of noise that gets in the signal path. One way to get rid of the noise is to introduce the transformers they used. One side of the transformer connects signal to signal and the other connects ground to ground. Behringer uses the same principle in their MicroHD HD400 hum eliminator (it works miracles by the way) so that could be the reason. The capacitors act as simple filters for high frequencies but their proper value is unknown. Here is what you do if you do not want to use the transformers:
Jump point A to point B and point C to point D. Do that for all the transformer spaces on that board. How good the final result will be is unknown. I like the input stage that I made but I am not happy with the output stage. Before we go to the output stage though we need to finish with the input stage. So, where does the signal from the small board goes? To the main sound board of course. That board requires modification too. Near the amplifier chip (LA4636) you will see two empty spaces for resistors, R203 and R204. I used 1000 ohm resistors but any value up to 4700 ohm would do. I also used two 10 uF non-polarized capacitors where the "100" indications are. That "100" means to use a jumper wire 10 mm long. The result is shown here:
Done, right? Nope. We need to supply the signal from the I/O board to those capacitors we just added. On the main sound board there is a connector labeled CN211, near connector CN209 where the signals from the CPU connect to. Connector CN211 is not present because the CDP100 is the lowest model but I am sure that it is used with the Privia series. There are two "75" spaces there that may or may not be populated with jumpers. If they are not populated you must solder jumpers as shown in the picture:
Then you need to work on points G, H and J. "G" is the ground signal, "H" is the left channel signal and "J" is the right channel signal. Three wires from those points must be connected to the small I/O board with the transformers. In a previous picture, letters "E" and "F" mark the spot. Letter "E" points to ground, where the wire from "G" must connect to. The pin next to "E" is the left channel input signal and that's where the wire from point "H" must connect to. Finally the third pin from "E" must connect to point J, the right input channel. In my mod I used a split-connector design so that I can remove one of the boards if needed while the other board stays inside the CDP.
Testing showed that these connections work well. There is however some fine print: if you attach cables to the input jacks and the device that connects on the other side of those cables is not "ON", you will hear some hum through the speakers of the CDP100. If you turn the input device "ON", the hum goes away. Before you send any sound from the input device to the CDP100 you need to lower the volume of that device to a minimum and then increase it carefully so that you do not saturate the amplifier of the CDP100. I keep the volume knob of the QY70 at 1/4 or less.
To top it I added a shield on the rear side of the I/O board that is located too close to the left speaker. A piece of aluminum foil laminated and cut in the proper shape was added and a ground wire was attached to it but it is not shown in the picture:
I will not discuss this part extensively because I am not happy with the end result. I designed a 2-stage output circuit with a gain of 3 or maybe slightly above that. I used the TL072 op-amp in the circuit. Each channel uses its own TL072 chip so that channel separation stays as high as possible. The input signal for that stage was taken right out of the volume potentiometer of the CDP100. The op-amps were biased at 6V because the power supply voltage is set at 12V by Casio. Electrolytic capacitors were used everywhere in order to filter out the DC. The circuit works but the high frequencies are muted when I connect the audio out to an external amplifier. It could be that the transformers I made or the capacitors I added in the I/O board create the problem. I did not check more on this issue, it's left for later on. An additional point to be considered here is that the CDP's speakers have to be muted with a headphone plug that connects to the headphone output if the modified line outputs are to be used. I think that the LA4636 can be muted with a DC signal somehow. But I had already spent much more time than expected on this project and I gave up. The two main targets had been reached anyway: better speakers and audio-in capability.
Finally we reach the ending of this long post. There is not much left to say. You can try anyone of the mods described here and if you mess up your keyboard it's your fault as always. The cost of materials for the mods was around $50 to $55, mainly because I buy components in bulk quantities so that individual parts come cheap. That cost does not include the speaker cloth replacement. If you try this your cost may be higher depending on where you buy the parts and how many you buy. Don't ask me to guide you through the process if you decide to work on it. I posted this for reference purposes only.
Enjoy your CDP100 anyway, with or without a mod.