Tag Archives: electronics

Emergency Room #9: Delta Labs Super Timeline ADM2048 Battery Burst

This is quite possibly the most interesting circuit I’ve worked on so far. I’ve always wondered how a digital delay works, and this late ’80s digital IC-based technology allowed the opportunity to reverse engineer the general concept. (I’ll share those findings in posts to come!) But before I could reverse engineer anything the unit had to be repaired to a working condition, as it had suffered the results of a failed board-mounted 4.6V battery:

Figure 1: Showing the Damage from Battery Leakage

There is an SRAM chip (NMC6504) used on this board to store 4 preset configurations (A, B, C & D), this is obviously being powered by the battery and was confirmed when I replaced the blown fuse, powered up the device, and no presets were able to be saved.

I couldn’t find the specs for this battery model online for a direct replacement, so I had to dig into many forums to find information and, eventually, a schematic. For legal reasons I won’t post the schematic here, contact me for more info if needed.

The schematics yield a 4.6V battery, and the test point for the battery voltage is labeled +5VB after the zener. I settled on a Varta 4.8V NiMH from JLS Batteries on eBay. It’s a through-hole component, but doesn’t match the footprint of the original battery, so I soldered wires from the + and – sides to the appropriate vias. To secure the wiring to the battery’s body I used electrical tape and twisted the wires together to provide additional support. The electrical tape may also help in preventing a large amount of damage to occur in the case that this NiMH fails in the future. A square piece of Velcro was used to secure the battery to the main board.

The use of a NiMH is based on the color of the corrosion that occurred on the copper traces and component leads. NiMH batteries contain the hydroxide anion (OH-). In the presence of an electrical current and physical contact, the hydroxide anion and copper can react to form copper (II) hydroxide, which is blue in color.

Copper (II) Hydroxide suspended in a Test-Tube

Copper (II) Hydroxide suspended in a Test-Tube (WikiMedia Commons)

To reiterate, this reaction is corrosive and literally eats away the copper traces on the board. If enough corrosion takes place the board may become irreparable! (Or at least extremely hard to repair without re-engineering and rebuilding the affected circuit). Luckily, this was not the case for this unit.

Once the replacement battery was installed I powered up the Super Timeline and presets were operational!  The contacts for the presets, however, were extremely unreliable. They would make contact maybe 60% of the time, so I ended up removing all of the preset button switching actuators. I tinned a thin, even layer of solder onto each contact and reinstalled them; that worked very well.

Figure 2: Programmable preset switches. The switch to far left has its' actuator removed. This is done by applying a precision flat-head screwdriver to the plastic lip between the lug sets.

Figure 2: Programmable preset switches. The switch to far left has its’ actuator removed. This is done by applying a precision flat-head screwdriver to the plastic lip between the lug sets.

After tinning the preset contacts I took aim to the copper (II) hydroxide build-up near the power and LFO circuits. I cleaned a toothbrush with distilled water and started brushing the corrosion build-up off the leads. I quickly noticed that the component labeling was getting scraped off as if they were old stickers. From that point on I was very careful in removing the corrosive material and mainly focused on relatively open areas between components. It’s not perfectly cleaned, but the unit was working and I didn’t want to ruin the labeling scheme, as it may prove useful down the road. (If anybody has an abrasive-free solution for cleaning copper (II) hydroxide from PCBs I’d be highly interested!) To remove the distilled water I used some combination of paper towels, a hair dryer, and an air compressor.


Thanks for reading! Notes on the circuit’s operation will definitely be a topic for future Talk Theory To Me posts. Stay tuned if you’re interested!

This fix was done for DMS Productions located in Ransomville, NY!

“With over 22 years experience in audio, 16 years experience in photography and over 5 years in video, DMS Productions has become a true multimedia company.”

Revived and Operational TC Electronics Finalizer

Emergency Room #3: T.C. Electronic Finalizer 96K Studio Mastering Processor

Recently I got my hands on a dead T.C. Finalizer Studio Mastering Processor from DMS Productions in Ransomville, NY. Right off the bat this sounds like a power supply problem.

The top panel is held down by 5 screws. Upon opening and inspecting the unit there are two main boards – the power supply unit and the main processing board. Focusing on the power supply, it is a switch-mode (SMPS) with a daughter-board vertically mounted next to the power transformer. SMPSs are inherently controlled via a pulse-width modulation (PWM) switching controller. For this unit, the PWM controller chip is mounted on the daughter-board – they use the Texas Instruments TL3842BP current mode PWM switching controller. The output of the PWM controller is fed into a TL431 Programmable Voltage Reference, which looks like a transistor at first glance. The voltage reference then is routed to other areas of the power supply to convert it to other DC voltages and clean it through some passive filtering.

PWM Controller Board Inside the Main Power Supply Unit

PWM Controller Board Inside the Main Power Supply Unit

Visual inspection reveals some corrosion on a capacitor and resistor sitting next to the daughter-board. There were also some other caps near the output power sources which had the same amount of corrosion.

Before electrical troubleshooting I thought it may be best to first search the forums on issues with this unit…boy were there issues! I’m notably grateful that Svart from GroupDIY was generous enough to post his fix in a thread describing almost this exact problem. He describes a bad electrolytic cap (100uF, rated 25V) next to the daughter-board being the cause of his issue..the same one that was found to be corroded on my unit. Immediately this capacitor was on the replacement list. To be safe, I also added the carbon film 1.2Mohm resistor sitting next to the bad cap as well as the other 82uF electrolytic caps near the output wiring which looked just as bad.

Showing the bad cap and resistor on the supply board

Showing the bad cap and resistor on the supply board

 

Having a list of the voltages carried on the output wires, along with having the labels of said voltages on the main board, I was ready to troubleshoot the board. I read zero volts on every rail against the proper ground/reference wires. The output was in fact dead. I turned to the PWM board and found out the output of the controller chip is pin 6. Measuring that on the oscilloscope revealed the chip wasn’t properly producing its PWM signal.

Highlighting the ground, Vcc, and output of the PWM interface connector

Highlighting the ground (1st pin), Vcc (2nd pin) and output (4th pin) of the PWM interface connector.

 

I also measured the Vcc power input for the PWM chip, which revealed about 14.36V. This particular controller has a minimum start-up supply voltage, which ranges between 14.5-17.5V. I believe this voltage is not being reached due to the damaged cap and resistor on the supply line to the PWM chip.

After replacing the caps and resistor I tested the output lines for proper voltages. The +15V and -15V rails were reading about +16V and -16V, respectively, which I was okay with. The only other DC voltage I was able to measure was the 6V line. The 5V lines were not readable, which was confusing at first. After some reading I found out it was a normal problem with another repair tech. There is apparently a “handshake” signal (carried by the brown wire) between the power supply and main board for activating the 5V line.

Even with that bit of advice, I still wasn’t sure if it was working. I hooked up the oscilloscope and found the PWM controller’s output pin to be operational at about 50kHz frequency.

Showing the output PWM signal on the o-scope

Showing the output PWM signal on the o-scope

With that reassurance, I connected the supply to the main board and the Finalizer arose from the dead! This was a fun one, for sure! It’s evident that this issue isn’t accidental or the fault of the consumer, as this same issue has occurred on multiple units. There must be a design flaw in the supply, maybe improper filtering, under-rated components, missed safeguards, etc. It would be a tremendous move for TC if they were to supply these replacements free of charge, if at all under proper warranty.

Emergency Room #2: Reviving a BOSS DS-1

Sometimes the problem is so simple, yet makes no sense how the issue arose in the first place…..

Let me explain.

I found an old BOSS DS-1 on Craiglist the other day. After a scenic 50 mile round-trip towards the Allegheny Plateau and back I finally had the chance to experience the classic distortion’s first impression. I plugged it in, flipped the standby switch, aaaaand…a LOUD hum with a subtle signal was present in the output. I had to investigate! (Needless to say, I was quite excited about this…one of the only instances where I’m not distraught about something not working.)

After gutting the DS-1 I realized a very simple issue…the output jack had no ground wire connecting the sleeve lug to the board’s ground. I couldn’t believe this, what a simple fix! This explains why I could still hear a faint signal coming through the wall of noise that the signal lug was picking up.

Notice no ground connection on the output (left) ground lug.

Notice no ground connection on the output (left) ground lug.

I ended up soldering a wire (green, below) from the ground lug to the board’s ground. The pedal works like new. For the board connection I chose the same solder joint by which the foot switch is connected, as that was the easiest & fastest connection point I could think of using.

Adding the ground connection (green). The black wire is the ground wire from the foot switch which terminates at the solder joint.

Adding the ground connection (green). The black wire is the ground wire from the foot switch which terminates at the solder joint.

But it begs the question….why?? I wonder if this issue got through the BOSS factory by accident? Did a previous repair technician forget to connect the ground lug back to the board? There were no signs of another technician touching it…a case for Sherlock if you ask me. Let me know your thoughts in the comments.

Inside Look: Rocktron’s Rampage Distortion Switching Circuit

Part 1: Important Components

1) ON Semiconductor MC14584BCP Hex Schmitt Trigger

An ideal inverter is one which, when taking in as input a logic HIGH (VDD), yields a logic LOW (0V) at the output and vice-versa. Now there are inverters…and then there are Schmitt Triggers. This isn’t the name of your typical high school punk band, it’s a device invented by Otto H. Schmitt in 1934 incorporating an electrical technique called hysteresis to the normal function of the inverter. At this point it’s enough to look at this component and think “inverter” but for fits and giggles let’s look at it a little closer.

Take a look at our device’s transfer characteristic below, taken from the MC14584B datasheet. The first dataset, the one with the upward facing arrow, describes what happens when the device switches from a higher input voltage to a lower input voltage. More specifically, when a decreasing voltage at the input crosses from logic HIGH through the value VT- and towards logic LOW, the device outputs a logic HIGH. Conversely, when the opposite happens (the input voltage increases from logic LOW through the value VT+) the device outputs a logic LOW. This is similar to how an ideal inverter works except the threshold values that determine the transfer characteristic are not equal. The difference in these thresholds (VT+ – VT-) is called the Hysteresis Voltage, which can be found in the electrical characteristics section of the datasheet. For this particular device the values that matter are: VT+ = 5.3V and VT- = 4.6V.

If you didn’t get that, no worries! It’s enough to think of these devices as simple inverters for our discussion. This particular package has 6 Schmitt Trigger devices built into it. In this particular switching circuit only 2 of them are used (pins 1, 2 and pins 12, 13). It also has double-diode protection on all inputs, which is important for the possibility of ESD trying to fry our chip.

Texas Instruments has a description of the operation of a Schmitt Trigger device: Understanding Schmitt Triggers

Here’s a YouTube video of YouTube’s very own Wez Lee explaining the operation of a Schmitt Trigger in the context of a comparator-based circuit.

Link to Mouser datasheet

2) Motorola MC14016BCP Quad Analog Switch/Quad Multiplexer

This particular analog/bilateral switch is comprised of three main I/O pins: IN, OUT, and CONTROL. Take a look at the block diagram to get a better picture here.

Let’s assume the IN is where your (analog) guitar signal is applied and the OUT is where you want to send your signal. When the CONTROL pin is logic LOW that signal is impeded and does not propagate through the device. However, when the CONTROL is logic HIGH your signal moves on to enjoy the adventures that await at the OUT terminal of the device.

That’s pretty much it! Oh, one important point to note is that unlike our Schmitt Trigger package this particular chip doesn’t have ESD protection on its I/Os. This is taken into account with the addition of double-diode protection in the design of our pedal’s circuit, as seen in the schematic.

Another thing that’s important to note is you don’t have to abide by the rules of INPUT or OUTPUT. With a bilateral analog switch the signal can propagate one way or the other. This will become important information when we look into the circuit’s operation.

Wikipedia has a good-enough description of analog(ue?) switches.

Link to datasheet

3) Motorola 2N5485 Small-Signal JFET

JFETs are commonly used in FX pedal activation controls (the classic Ibanez Tube Screamer is the first that comes to mind). In this particular circuit all you need to know is that its used as a switch. For those not familiar with the design of a JFET, take a look at the pinout below. Current can flow from the Drain to the Source terminals when a sufficient voltage is applied to the Gate terminal. If the amount of voltage is not sufficient current does not flow. Sounds like a switch, right? We’ll see how and why its used in the circuit later.

For more information on JFETs check out this ElectronicsTutorials article.

Part 2: Schematic Overview

In the following schematics the green highlights logic HIGH signals and the red signifies the analog guitar signal.

Effect DISENGAGED:

Let’s first consider the control circuit when the effect is disengaged (i.e. when the jumper terminals J2-1 and J2-2 are left open). The power source’s 9V is seen at the input (pin 13) of one of the Schmitt Trigger inverters. A logic LOW is seen at the output (pin 12) which is also connected to the LED (which is switched OFF because of insufficient current). This LOW signal is also seen at the input of the other Schmitt Trigger inverter (pin 1), driving its output (pin 2) to a logic HIGH. This logic HIGH performs two actions:

  1. It provides sufficient voltage to the gate of the JFET device so the clipper circuit can be bypassed. My guess is that this is to ensure that potentially amplified noise cannot be coupled or propagated to the rest of the circuit…your thoughts are valuable here!
  2. It sends the CONTROL input (pin 13) of the analog switch to a logic HIGH, allowing any analog signal to pass through. In this case the signal comes from the input of the pedal, is buffered by the 5532 op amp circuit, propagates through the input (pin 1) and output (pin 2) of the switch, and gets sent to the output of the pedal.

Effect ENGAGED:

Now let’s start the party by engaging our effect into action, thereby connecting J2’s terminals. In this case current from the 9V source is shunted to ground through 1R28, forcing a logic LOW at the input (pin 13) of the Schmitt Tigger. The resulting HIGH signal at the output (pin 12) turns on the LED, ceases the bypassing of the clipper circuit (by no longer supplying the JFET’s gate with sufficient voltage), and closes the other analog switch, allowing the processed signal to flow from the filter stage of the circuit to the output of the pedal.

Part 3: The Inside Look

I like to be comprehensive when it comes to showing what I’ve done experimentally. The rest of this section is nothing more than a listing of what I’ve captured from the relevant zones of the pedal’s circuitry, in this case the switch control circuit.

DC Voltage Values:

The following DC values were captured when the effect was disengaged:

And here, with the pedal engaged:

Waveforms:

The following waveforms were captured when the effect was disengaged:

Pin 1, MC140146B (Disengaged)

Pin 2, MC14016B (Disengaged)

Pin 10, MC14016B (Disengaged)

Pin 11, MC14016B (Disengaged)

Gate, 2N5485 (Disengaged)

Source, 2N5485 (Disengaged)

Drain, 2N5485 (Disengaged)

And the following, with the effect engaged:

Pin 1, MC14016B (Engaged)

Pin 2, MC14016B (Engaged)

Pin 11, MC14016B (Engaged)

Gate, 2N5485 (Engaged)

Source, 2N5485 (Engaged)

Drain, 2N5485 (Engaged)

Reference Datasheets:

MC14584B Hex Schmitt Trigger

MC14016BCP Quad Analog Switch/Quad Multiplexer

2N5485 JFET Amplifier