Great observation, since these batteries are rated at 8.4 V nominal, they typically will output between about 9.5 V (fully charged) and about 7.5-8V (on low charge), so if I keep an eye on them and check their voltage regularly there should be no problem, this is only a prototype, on the next version (if I use batteries) maybe I will include a low battery indicator and perhaps also a shutdown feature, to turn off the device when the voltage is too low for operation.
I usually don't use tantalum capacitors, so I have never seen one explode, I hope that I will continue this way. 🤞
I have actually exploded one small electrolytic capacitor (10 uF or so) by reversing it's polarity, but it only popped on the underside and some electrolyte came out.
One colleague of the faculty had 2nd degree burns on his arm due to an exploding capacitor, it was an SMD power supply reservoir capacitor included on a Spartan 3E FPGA development kit. Presumably it was a cheap capacitor affected by the "plague" (i. e. incorrect electrolyte formulation).
Yes, everyone should, the manufacturers always state that in manuals and datasheets.@danny92 That is why I always, really always wear protection glasses while working with exposed electronic boards
I never use protection gear, but I know I should.
A post is missing between #7,146 and #7,147, maybe the post didn't went thru because I was offline.
Even with the LED and resistor, the +5 V regulator won't work, only if I connect the batteries in the order I've mentioned in the last post. I'm not sure if it's the regulator or something to do with the AD1862s, but the voltage at the +Vd pins and at the +5 V regulator is always about -0.7 V when the power supplies fail, and about 4.95 V when it's working.
Once the power supply it's ok everything works fine. 😁
This setup AD1862+AD811 is the best I've listened so far, it's odd, but everything digital that I've listened in my life appear to be much less natural sounding. This one sounds like a very well recorded/printed vinyl record, it doesn't seem digital at all. The soundstage is also great, giving the sensation of actually being there. Every instrument seems to be rendered almost to perfection.😆
The AD1866s sounds good and the AD1851J+LM6171 holds currently the second place in the DAC prototypes that I've made so far, but this is in another level. 🙃
Even with the LED and resistor, the +5 V regulator won't work, only if I connect the batteries in the order I've mentioned in the last post. I'm not sure if it's the regulator or something to do with the AD1862s, but the voltage at the +Vd pins and at the +5 V regulator is always about -0.7 V when the power supplies fail, and about 4.95 V when it's working.
Once the power supply it's ok everything works fine. 😁
This setup AD1862+AD811 is the best I've listened so far, it's odd, but everything digital that I've listened in my life appear to be much less natural sounding. This one sounds like a very well recorded/printed vinyl record, it doesn't seem digital at all. The soundstage is also great, giving the sensation of actually being there. Every instrument seems to be rendered almost to perfection.😆
The AD1866s sounds good and the AD1851J+LM6171 holds currently the second place in the DAC prototypes that I've made so far, but this is in another level. 🙃
Last edited:
Hmmm. Then it very well could be the faulty regulator. If you can, order onsemi 7805, it's quite cheap. This way we can narrow it down to batteries if it shows the same symptoms. The ad chip itself shouldn't be at fault here as you have a load on the regulator right in front of it.
For 78XX, 79XX, LM317/337 and similar, at least 4V input/output differential voltage is recommended. It should be taken into account that sometimes there may be a voltage drop in the network, and ripple voltage is also involved. I usually practice 5-6V differences, and it's not bad to put an initial load on the regulator of some 10mA to make it work properly. For the LDO, 2-3V is enough to work safely, although again there is always a possible voltage drop and ripple voltage.
Looks like some protection in the regulator is kicking in. They have overcurrent and thermal protection. Check all wires, voltages in front and behind the regulator, regulator heating, and if everything is fine, replace the regulator.
And frankly, the AD1862 deserves slightly better regulators than the 78/79XX.
I've found in my stock a Fairchild LM7805AC, so I promptly replaced the L7805CV. 😁
Now the +5 V is always stable and never seems to fail at startup, the LED always shines, I can power up in any sequence, the only thing that I must pay attention is that -Vs (the -12 V supply) must be applied first, like is mentioned at the AD1862s datasheet.
I thought ST and DigiKey were reliable brands, maybe I was wrong
Do you think I can ask for a refund?
Now I only have one ST regulator in my setup, an L7912ACV (high grade, I guess), the two others are from ON Semi, MC7905B and MC7812B.
Now the +5 V is always stable and never seems to fail at startup, the LED always shines, I can power up in any sequence, the only thing that I must pay attention is that -Vs (the -12 V supply) must be applied first, like is mentioned at the AD1862s datasheet.
I thought ST and DigiKey were reliable brands, maybe I was wrong
Do you think I can ask for a refund?
Now I only have one ST regulator in my setup, an L7912ACV (high grade, I guess), the two others are from ON Semi, MC7905B and MC7812B.
Last edited:
Great!
Do note that even the best of brands can have a number % of devices failures, it is a series production after all and not each and any component goes trough qa, but rather as several random specimen from a batch. ST has been good to me, so no need to blame it as a whole
Yes you can always go trough rma, after all you received a faulty component. Not so sure if it's worth the time for a <1$ component
Do note that even the best of brands can have a number % of devices failures, it is a series production after all and not each and any component goes trough qa, but rather as several random specimen from a batch. ST has been good to me, so no need to blame it as a whole
Yes you can always go trough rma, after all you received a faulty component. Not so sure if it's worth the time for a <1$ component
Yes, you're right, I also have some ST components that are working reliably, some for a long time.
Before replacing the L7805 with the LM7805, I've tried to replace it by another L7805 from the same tube and it also failed. It seems like this batch has some problems with the protection circuitry, that seems to be overreacting and shutting down the regulator for apparently no reason, the batteries are fine, with an output voltage over 9 V, I don't know if the -0.7 V come from the regulator output or from some protection inside the AD1862s.
I have a L7805 from this batch powering a AD1851J + LM6171 prototype and it is working reliably, so maybe the protection only kicks in under certain circumstances which I don't know exactly which.
Before replacing the L7805 with the LM7805, I've tried to replace it by another L7805 from the same tube and it also failed. It seems like this batch has some problems with the protection circuitry, that seems to be overreacting and shutting down the regulator for apparently no reason, the batteries are fine, with an output voltage over 9 V, I don't know if the -0.7 V come from the regulator output or from some protection inside the AD1862s.
I have a L7805 from this batch powering a AD1851J + LM6171 prototype and it is working reliably, so maybe the protection only kicks in under certain circumstances which I don't know exactly which.
AD1862 has not specified power up sequences.
"... the limitation that –VL may not be more negative than –VS." it is meant for normal operation. Inside the dac are switches with the ability to distinguish TTL or CMOS logic and I think it is specified for that reason (for normal operation, not as a limitation against chip damage or bad power-up)
You can study all that from AD1862 related patents: U.S. Patent Numbers: 4,349,811; 4,857,862; 4,855,618; 3,961,326; 4,141,004; 4,902,959.