max3490esa_MAX3490EESA+

AVAILABLE

Functional Diagrams

Pin Configurations appear at end of data sheet.

Functional Diagrams continued at end of data sheet.

UCSP is a trademark of Maxim Integrated Products, Inc.

For pricing, delivery, and ordering information, please contact Maxim Direct

at 1-888-629-4642, or visit Maxim’s website at www.maximintegrated.com.

_______________General Description

The MAX3483, MAX3485, MAX3486, MAX3488,

MAX3490, and MAX3491 are 3.3V, low-power trans-

ceivers for RS-485 and RS-422 communication. Each

part contains one driver and one receiver. The

MAX3483 and MAX3488 feature slew-rate-limited dri-

vers that minimize EMI and reduce reflections caused

by improperly terminated cables, allowing error-free

data transmission at data rates up to 250kbps. The par-

tially slew-rate-limited MAX3486 transmits up to

2.5Mbps. The MAX3485, MAX3490, and MAX3491

transmit at up to 10Mbps.

Drivers are short-circuit current limited and are protect-

ed against excessive power dissipation by thermal

shutdown circuitry that places the driver outputs into a

high-impedance state. The receiver input has a fail-safe

feature that guarantees a logic-high output if both

inputs are open circuit.

The MAX3488, MAX3490, and MAX3491 feature full-

duplex communication, while the MAX3483, MAX3485,

and MAX3486 are designed for half-duplex communi-

cation.

________________________Applications

Low-Power RS-485/RS-422 Transceivers

Telecommunications

Transceivers for EMI-Sensitive Applications

Industrial-Control Local Area Networks

____________________________Features

o

Operate from a Single 3.3V Supply—

No Charge Pump!

o

Interoperable with +5V Logic

o

8ns Max Skew (MAX3485/MAX3490/MAX3491)

o

Slew-Rate Limited for Errorless Data Transmission

(MAX3483/MAX3488)

o

2nA Low-Current Shutdown Mode

(MAX3483/MAX3485/MAX3486/MAX3491)

o

-7V to +12V Common-Mode Input Voltage Range

o

Allows up to 32 Transceivers on the Bus

o

Full-Duplex and Half-Duplex Versions Available

o

Industry Standard 75176 Pinout

(MAX3483/MAX3485/MAX3486)

o

Current-Limiting and Thermal Shutdown for

Driver Overload Protection

______________Ordering Information

Ordering Information continued at end of data sheet.

* Contact factory for for dice specifications.

3.3V-Powered, 10Mbps and Slew-Rate-Limited

True RS-485/RS-422 Transceivers

19-0333; Rev 0; 12/94

PART

MAX3483CPA

MAX3483CSA

MAX3483C/D

0°C to +70°C

TEMP. RANGE

PIN-PACKAGE

8 Plastic DIP

8 SO

Dice*

PART

NUMBER

GUARANTEED

DATA RATE

(Mbps)

SUPPLY

VOLTAGE

(V)

HALF/FULL

DUPLEX

SLEW-RATE

LIMITED

DRIVER/

RECEIVER

ENABLE

SHUTDOWN

CURRENT

(nA)

PIN

COUNT

MAX3483

0.25

Half

Yes

2

8

MAX3485

10

Half

No

Yes

2

8

MAX3486

2.5

Half

Yes

2

8

MAX3488

0.25

Full

Yes

No

8

MAX3490

10

Full

No

8

MAX3491

10

3.0 to 3.6

Full

No

Yes

2

14

______________________________________________________________Selection Table

MAX3483EPA

MAX3483ESA

-40°C to +85°C

8 Plastic DIP

8 SO

MAX3485CPA

MAX3485CSA

MAX3485C/D

0°C to +70°C

8 Plastic DIP

8 SO

Dice*

MAX3485EPA

MAX3485ESA

-40°C to +85°C

8 Plastic DIP

8 SO

MAX3483/MAX3485/MAX3486/

MAX3488/MAX3490/MAX3491

  • 0
    点赞
  • 0
    收藏
    觉得还不错? 一键收藏
  • 0
    评论
ESA_block是一种自注意力机制,可以应用于自然语言处理中的文本表示学习。下面是一个基于Keras实现ESA_block的示例代码: ```python import tensorflow as tf from tensorflow.keras import layers class ESA_block(layers.Layer): def __init__(self, embed_dim, num_heads, dense_dim, rate=0.1): super(ESA_block, self).__init__() self.embed_dim = embed_dim self.num_heads = num_heads self.dense_dim = dense_dim self.rate = rate self.dense = layers.Dense(dense_dim, activation='relu') self.dropout1 = layers.Dropout(rate) self.dense2 = layers.Dense(embed_dim) self.dropout2 = layers.Dropout(rate) self.norm1 = layers.LayerNormalization(epsilon=1e-6) self.norm2 = layers.LayerNormalization(epsilon=1e-6) self.attention = layers.MultiHeadAttention(num_heads=num_heads, key_dim=embed_dim) def call(self, inputs, training=True): x = inputs attention_out = self.attention(x, x) attention_out = self.dropout1(attention_out, training=training) out1 = self.norm1(x + attention_out) dense_out = self.dense(out1) dense_out = self.dropout2(dense_out, training=training) out2 = self.norm2(out1 + self.dense2(dense_out)) return out2 ``` 在上述代码中,我们定义了一个名为ESA_block的自定义Keras层。这个层包括了一个多头注意力机制和两个前馈全连接层。该层的输入是一个形状为(batch_size, seq_len, embed_dim)的张量,其中seq_len表示序列长度,embed_dim表示嵌入维度。输出也是一个形状为(batch_size, seq_len, embed_dim)的张量。 在call方法中,我们首先通过调用注意力机制来计算自注意力向量。然后我们添加一个残差连接和一个层归一化操作。接下来,我们将注意力向量通过两个前馈全连接层,然后再次添加残差连接和层归一化操作。最后,我们返回输出张量。 要使用该层,可以像使用Keras中的其他层一样添加到自己的模型中。例如: ```python inputs = layers.Input(shape=(seq_len, embed_dim)) x = ESA_block(embed_dim, num_heads, dense_dim)(inputs) ... ``` 这里我们定义了一个输入张量,然后将它传递给ESA_block层,最后将输出张量传递给下一个层。
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值